Search

Recently, computer vision and artificial intelligence are being used as enabling technologies for plant phenotyping studies, since they allow the analysis of large amounts of data gathered by the sensors. Plant phenotyping studies can be devoted to the evaluation of complex plant traits either on the aerial part of the plant as well as on the underground part, to extract meaningful information about the growth, development, tolerance, or resistance of the plant itself. All plant traits should be evaluated automatically and quantitatively measured in a non-destructive way. This paper describes a novel approach for identifying plant roots from images of the root system architecture using a convolutional neural network (CNN) that operates on small image patches calculating the probability that the center point of the patch is a root pixel. The underlying idea is that the CNN model should embed as much information as possible about the variability of the patches that can show chaotic and heterogeneous backgrounds. Results on a real dataset demonstrate the feasibility of the proposed approach, as it overcomes the current state of the art. • Root systems must be monitored to assess the growth and well-being of a plant. • State-of-the-art approaches mainly use classic ML or U-networks for segmentation. • CNNs can be used for monitoring considering patch-based information. • These models are simpler and faster, and provide better segmentation performance. [ABSTRACT FROM AUTHOR]

Nowadays, for controlling crime, surveillance cameras are typically installed in all public places to ensure urban safety and security. However, automating Human Activity Recognition (HAR) using computer vision techniques faces several challenges such as lowlighting, complex spatiotemporal features, clutter backgrounds, and inefficient utilization of surveillance system resources. Existing attempts in HAR designed straightforward networks by analyzing either spatial or motion patterns resulting in limited performance while the dual streams methods are entirely based on Convolutional Neural Networks (CNN) that are inadequate to learning the long-range temporal information for HAR. To overcome the above-mentioned challenges, this paper proposes an optimized dual stream framework for HAR which mainly consists of three steps. First, a shots segmentation module is introduced in the proposed framework to efficiently utilize the surveillance system resources by enhancing the lowlight video stream and then it detects salient video frames that consist of human. This module is trained on our own challenging Lowlight Human Surveillance Dataset (LHSD) which consists of both normal and different levels of lowlighting data to recognize humans in complex uncertain environments. Next, to learn HAR from both contextual and motion information, a dual stream approach is used in the feature extraction. In the first stream, it freezes the learned weights of the backbone Vision Transformer (ViT) B-16 model to select the discriminative contextual information. In the second stream, ViT features are then fused with the intermediate encoder layers of FlowNet2 model for optical flow to extract a robust motion feature vector. Finally, a two stream Parallel Bidirectional Long Short-Term Memory (PBiLSTM) is proposed for sequence learning to capture the global semantics of activities, followed by Dual Stream Multi-Head Attention (DSMHA) with a late fusion strategy to optimize the huge features vector for accurate HAR. To assess the strength of the proposed framework, extensive empirical results are conducted on real-world surveillance scenarios and various benchmark HAR datasets that achieve 78.6285%, 96.0151%, and 98.875% accuracies on HMDB51, UCF101, and YouTube Action, respectively. Our results show that the proposed strategy outperforms State-of-the-Art (SOTA) methods. The proposed framework gives superior performance in HAR, providing accurate and reliable recognition of human activities in surveillance systems. [ABSTRACT FROM AUTHOR]

Traffic congestion has become a critical issue in metropolitan areas, inflicting significant hardships on individuals on a daily basis. This adverse situation poses multifaceted challenges, including increased air pollution, heightened fuel consumption, and compromised road safety. Deep learning algorithms, encompassing both Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN) have showcased remarkable performance in areas such as computer vision, natural language processing, healthcare, finance, and autonomous systems. This research paper proposes a novel methodology to develop a vehicle detection model - VehiNet, which can subsequently be refined to create an automated traffic signal system. The study explores the effectiveness of various CNN, RNN, and deep learning algorithms and techniques in accomplishing this objective. The dataset picked for the study in addition to being compiled from internet resources, was also painstakingly assembled manually using a variety of methods to collect over 2000 photographs from various angles, lighting conditions, and locations within Bengaluru city. The main motive is to apply CNN, RNN, and other Deep Learning classification techniques for object detection and classification of Indian vehicles and see how well these deep learning algorithms and techniques affect prediction efficiency. This research stands out from others and makes its novelty clear thanks to the carefully collected dataset and the detailed analysis of the methodologies. [ABSTRACT FROM AUTHOR]

The integration of deep learning, computer vision, and advanced algorithms has ushered in a transformative era in the prediction of human driving behavior, consequently revolutionizing road safety. This paper focuses on an innovative convergence of technology that addresses critical issues like driver fatigue and distracted driving by automatically identifying and categorizing diverse driving behaviors. Neural network architectures, such as VGG16, AlexNet, and ResNet are described in this paper that have propelled accuracy in behavior classification to remarkable levels. However, the quest for safer roads remains ongoing, with promising avenues lying ahead. First and foremost, the creation of extensive, diverse, and meticulously annotated datasets is paramount. These datasets serve as the bedrock upon which future models can be trained, enhancing their robustness and generalizability across a spectrum of driving scenarios. Real-time models represent another pivotal frontier. These models hold the potential to provide timely interventions and support systems for drivers, thus preventing accidents proactively. The exploration of hybrid techniques that amalgamate the strengths of various neural network architectures presents an exciting avenue, promising to further push the boundaries of prediction accuracy. Furthermore, this paper also discusses the fusion of multi-modal data, encompassing sensor data from IoT and smartphone devices, that holds immense promise. This holistic approach promises a more comprehensive understanding of driver behavior by integrating diverse data sources, ultimately contributing to the creation of safer road environments.In this research paper, we also explore these cutting-edge developments in deep learning and computer vision, emphasizing technical novelty and innovation. Through an interdisciplinary approach, we envision a future where the synergy of technology, data, and human behavior leads to a substantial reduction in road accidents and improved road safety. [ABSTRACT FROM AUTHOR]

The proliferation of fake images in today's digital landscape poses a significant threat to various domains, including media integrity, social media, and online security. Recognizing the urgent need to distinguish real images from their deceptive counterparts, this paper underscores the importance of developing a robust detection system. While substantial efforts have been made in the realms of computer vision and deep learning, the advent of Generative Adversarial Networks (GANs) has added a new layer of complexity to this challenge. In response to these evolving threats, we present a novel two-step methodology for detecting fake images, with a specific focus on those generated by GANs. Our approach harnesses the combined strengths of GANs and traditional Convolutional Neural Networks (CNNs), offering a comprehensive solution that significantly enhances accuracy in identifying both fake images and machine-generated fake images. The results of our experiments demonstrate the efficacy of our methodology. Using CNNs alone, we achieved a training accuracy of 87%. However, when employing the collaborative power of GANs and CNNs, our model exhibited a remarkable accuracy rate of 94.4%. This substantial improvement underscores the superiority of the GANs+CNN approach, suggesting its potential as a groundbreaking solution in the realm of fake image detection. This research opens up new horizons in fields such as media forensics, social media monitoring, and online security, where the ability to discern genuine content from manipulated or synthetic media is of paramount importance. The promising outcomes of this study not only provide an immediate and effective solution but also pave the way for further exploration and innovation in this critical area of digital security. [ABSTRACT FROM AUTHOR]

The image-based automatic plant disease identification has emerged as a prominent research area, driven by recent advancements in machine learning algorithms. Conventional disease detection methods rely on manual feature extraction from acquired images, involving extensive expertise from experts and pathologists. This labor-intensive practice is often time-consuming and impractical. With the recent advancements in digital technologies and computer vision, Deep neural networks address these challenges by learning the features automatically. This paper focuses on the EfficientNet B0 architecture-based deep neural network model integrated with a convolutional block attention mechanism module for identifying major wheat crop diseases. The attention module in the EfficientNet B0 model improves its performance by emphasizing the disease-infected regions in leaf images. To evaluate the proposed methodology, experiments were conducted on the WheatRust21 image dataset, comprising 6556 images of three major fungal diseases, namely, stripe rust, leaf rust, and stem rust, collected in natural field conditions from 2019 to 2021. Our developed EfficientNet B0-CBAM model performed well with a testing accuracy of 98.70% and an impressive F1 score of 98% on the dataset. The Grad-CAM visualization demonstrates that the attention module focuses more on the infected region and plays a crucial role in wheat disease identification. The results are quantitatively substantial, with the model achieving high accuracy and other evaluation metrics scores with an average of 98.5 %. These findings hold promise for practical applications in agriculture and sustainable crop management. [ABSTRACT FROM AUTHOR]

The complex flow of a supersonic jet (Mach number 2.4–2.5) around the blunt aerodynamic body (consisting of a cylinder with a diameter of 16 mm and a truncated cone) was studied using the high-speed shadowgraph technique and computer vision-based visual data processing. The problem considered is directly related to the rocket and space systems stages safe separation problem in a sufficiently dense atmosphere. For systems with a sequential arrangement of stages, their separation often occurs under the rocket engine jet action. The occurrence of lateral forces in violation of axial symmetry can lead to overturning of the discarded stage and undesirable interaction with the outgoing stage. The goals of the paper were to study the pulsation characteristics of the incoming supersonic jet streamlining the body, as well as the pulsations of the bow shock wave in both axisymmetric and asymmetric flows. The recording frame rate in the experiments was up to 775 000 fps. Computer vision and various machine learning approaches were used to process a large number of shadowgraph images. Canny edge detection and Hough transform, as well as a convolutional neural network (CNN) based on YOLOv2 architecture, were used to automatically track the position of the bow shock relative to the body. Jet sound generation was studied - the amplitude dependence on the distance from the nozzle and the evolution in time. The position of the bow shock relative to the streamlined body and flow pulsations over time were analyzed. Bow shock pulsation spectra at zero angle of attack were compared to those at a small angle of attack (α = 2.5°–3°). The power spectral density was also estimated. The slope of the spectra for the zero angle of attack case was found to be horizontal and the spectra for the small angle of attack case had a slope close to −5/3. The strongest of the measured bow shock frequencies were found to be close to those of the jet mixing layer pulsations. The hardware signal spectrum was also measured and this signal can be considered as noise. The described approach makes it possible to obtain the physical characteristics of the flow at any point of interest. It was also shown that the automation of flow visualization experiments using computer vision techniques allows to increase the speed of data processing and obtaining new physical information, which may be important for engineers developing aircraft and spacecraft technologies. • Supersonic jet flow around a blunt body was studied experimentally and also numerically. • Computer vision tools were applied to supersonic flow investigation. • Spectra of jet generated acoustic waves were determined. • Time dependent bow shock positions were automatically measured by the neural network. • Amplitudes of bow shock pulsations and flow fluctuations depend on the angle of the body relative to the jet direction. [ABSTRACT FROM AUTHOR]

In recent years, the advancement of deep learning technologies has significantly impacted various domains and the field of transportation is no exception. As the demand for robust and accurate traffic sign classification systems continues to rise, this study presents an in-depth exploration and comparison of various techniques employed in the field. Focusing on state-of-the-art methodologies, we assess the effectiveness and performance of multiple classification approaches for traffic sign recognition. The review presents an extensive survey of the literature, encompassing traditional computer vision methods, machine learning algorithms, and the latest advancements in deep learning. The comparative analysis aims to identify the strengths and limitations of each technique, considering factors such as computational efficiency and accuracy. Additionally, the paper implements four models—CNN, ResNet50, VGG19, and EfficientNetB7—for traffic sign classification on the GTSRB dataset, the accuracy results are reported as 99.25%, 99.28%, 98.95%, and 98.24% respectively. [ABSTRACT FROM AUTHOR]

Image registration is a fundamental problem in computer vision and robotics. Recently, learning-based image registration methods have made great progress. However, these methods are sensitive to abnormal transformation and have insufficient robustness, which leads to more mismatched points in the actual environment. In this paper, we propose a new registration framework based on ensemble learning and dynamic adaptive kernel. Specifically, we first use a dynamic adaptive kernel to extract deep features at the coarse level to guide fine-level registration. Then we added an adaptive feature pyramid network based on the integrated learning principle to realize the fine-level feature extraction. Through different scale, receptive fields, not only the local geometric information of each point is considered, but also its low texture information at the pixel level is considered. According to the actual registration environment, fine features are adaptively obtained to reduce the sensitivity of the model to abnormal transformation. We use the global receptive field provided in the transformer to obtain feature descriptors based on these two levels. In addition, we use the cosine loss directly defined on the corresponding relationship to train the network and balance the samples, to achieve feature point registration based on the corresponding relationship. Extensive experiments on object-level and scene-level datasets show that the proposed method outperforms existing state-of-the-art techniques by a large margin. More critically, it has the best generalization ability in unknown scenes with different sensor modes. [ABSTRACT FROM AUTHOR]

Age estimation from facial images is a challenging topic in computer vision since it can automatically label the human face with an exact age according to various physical or biological characteristics, such as facial structure, spots, and wrinkles. Additionally, it has substantial applications in many fields, including healthcare, security, entertainment, and education. There are a lot of techniques to estimate age, but the most popular one is the convolutional neural network (CNN), which offers high accuracy but needs a lot of training time and much more labeled data to achieve it. Another popular technique is transfer learning with feature extraction, which can provide higher accuracy with faster training time because it uses a pre-trained model rather than building one from scratch but has limited flexibility and a substantial risk of overfitting. Due to the advantages and disadvantages of both methods, this study analyzes and compares the effectiveness of transfer learning and CNN methods for age estimation from facial images. Using R-squared and RMSE as test metrics, the results show that the CNN method offers a better process and accuracy than the transfer learning method, with a higher R-squared value and a lower RMSE value. [ABSTRACT FROM AUTHOR]

Automated detection of pineapple-picking robots in complex agricultural environments is challenging due to uneven lighting and occluded fruit. In this paper, a lightweight pineapple detection model based on the YOLOv7-tiny model is presented for real-time accurate detection by agricultural robots. Firstly, The SIoU loss function is designed to substitute for the CIoU loss function in the initial model. The mismatched direction can be used to minimize the overall degree of freedom and expedite model convergence. Then, incorporating the CBAM module into the backbone network enhances the model's capacity to emphasize key features of the pineapple fruit, resulting in improved generalization ability and overall robustness. Ultimately, the improved YOLOv7-tiny model achieves a mAP@0.5 of 96.9%, surpassing the original model by 1.6%. [ABSTRACT FROM AUTHOR]

The segmentation of glioma by computer vision is one of the hot topics in medical image analysis, which further helps doctors to make a better treatment plan for glioma. At present, convolutional neural networks (CNN) with multi-kernels are the mainstream method to identify glioma regions. However, the segmentation is strongly affected if the intensity dissimilarity between adjacent glioma regions is small. To solve this challenge, we propose an attention-based multimodal glioma segmentation with multi-attention layers for small-intensity dissimilarity to focus on the small-intensity dissimilarity glioma regions. Firstly, to reduce background interferences, this paper proposes data enhancement in glioma-centered regions. In addition, a random multi-dimensional data view is generated in the glioma regions to reduce overfitting. Secondly, we embed a 3D U-Net to proposed attention layers, which focus on the intensity dissimilarity between adjacent glioma regions, and adaptively mine the glioma-related features, solving the problem that the existing algorithms are insensitive to the small-intensity dissimilarity between adjacent glioma regions. In particular, each attention layer can adaptively highlight valuable glioma features and suppress unrelated ones. Finally, experiment results on the multimodal brain tumor segmentation challenge (BraTS) 2020 dataset validate the effectiveness of the proposed method, where the Dice Similarity Coefficients (DSC) are improved on the segmentation of whole tumor (WT), tumor core (TC), and enhanced tumor (ET) regions, reaching higher results at 0.7803, 0.8831 and 0.8172 for WT, TC and ET regions respectively. Also, we make a test on the public dataset BraTS2019, reaching the results at 0.7675, 0.8925, and 0.8110 for WT, TC, and ET regions, respectively. [ABSTRACT FROM AUTHOR]

Mobile assistance systems in demanding wind turbine maintenance processes can help to reduce the work time of individual tasks. One such task is the measurement of carbon brushes, which need replacement when mechanical abrasions occur. We developed an image-based detection of different types of carbon brushes using an autoencoder network. After the detection, we adjusted the alignment of each carbon brush to get the height to length ratio, without additional depth information, for a reasonable estimation of the mechanical abrasion and a suggestion for a possible replacement. This leads to a more efficient workflow and less manual work during maintenance procedures. [ABSTRACT FROM AUTHOR]

With the advent of deep learning in recent years, the integration of computer vision with natural language processing has garnered a lot of attention. Generating descriptions from images is one of the most intriguing and focused areas of Machine Learning that faces a number of obstacles, particularly when describing images in languages other than English. In image captioning, a computer is trained to understand the visual information of an image and then generate a description based on the image features and reference sentences. Having an application that automatically describes events in their environment and then translates them into a caption or message can make a significant contribution to society. This paper presents a multi-layered CNN-LSTM neural network model that is utilized to recognize and generate Hindi captions for the objects in images. In addition, a variety of models were trained by adjusting hyperparameters and the number of hidden layers to find the optimum model and maximize the likelihood of the resultant Hindi description. Moreover, after testing the effectiveness of our models, it was observed that our model has shown an increase of 34.64% and 29.13% in terms of BLEU score (Unigram) and BLEU score (Bigram) respectively when compared to the existing work done in this field. Image captioning in Hindi can have a multitude of applications in today's society, and it can also provide a user-friendly interface for Hindi speakers. [ABSTRACT FROM AUTHOR]

Due to the intensive use of the domestic refrigerator based on vapor compression and the particular design of the no-frost type, it is essential to propose economic and light strategies to optimize the refrigerator's energy consumption. When a large amount of frost accumulates on the evaporator, heat exchange decreases, affecting the cooling capacity. The amount of frost accumulation on the evaporator surface plays a critical role in optimizing frost removal. For this reason, this paper presents a novel method that uses visual information to classify the frost accumulated on the evaporator surface accurately. This method has two parts: a data acquisition system based on an RGB sensor and an intelligent model that processes the visual information and classifies the frost accumulated on the evaporator surface. The data acquisition system creates the datasets for the training and validation of the intelligent model. This smart model is based on convolutional neural networks. Many tests were performed on a pair of different domestic refrigerators using typical usage habits for data collection. Each test was performed continuously for 24 h, emulating real-life usage. One of the main contributions of this work is a new, accurate, and low-cost method to classify the amount of frost accumulated on the evaporator surface. Furthermore, one advantage of the proposed method is the ease of integrating into any domestic refrigerator. The results indicate that this proposed method can classify with an accuracy as high as 99.75 % four different levels of frost accumulated on the evaporator surface of a domestic refrigerator. [ABSTRACT FROM AUTHOR]

Urban pollution is a massive global problem, especially in industrialized and developing nations. Visual pollution is an issue concerned with the external noticeable appearance of the modern urban areas causing human health disorders, emotional distress, driving distraction, environmental hazards, etc. Amidst the plethora of different forms of environmental pollution, visual pollution deteriorates the aesthetics of an urban environment, endorsing the importance of research and assessing it from different dimensions. The main objective of this study is to initialize a new concept of automatic identification and classification of visible contaminants related to textile industries using computer vision techniques. In this work, deep learning techniques have been applied for the automatic detection and classification of three categories of textile-based visual pollutants, i.e., cloth garbage, advertising billboards and signages, and textile dyeing waste materials. Initially, 1,709 visual pollutants images were obtained through web crawling of search engines. Additionally, 954 images were collected from two local garments factories, roadside vendors and shopping malls of Bangladesh. Next, the dataset was manually annotated by an open-source labeling tool. Finally, various deep learning techniques, Faster R-CNN, YOLOv5, and EfficientDet, have been used to classify the obtained dataset automatically. The EfficientDet framework achieved the best performance with 97 % and 93 % training and test accuracies, respectively. The YOLOv5 approach exhibits acceptable precision with a considerably lower number of epochs. The proposed automated classification system is expected to create future visual pollution ratings for the textile industries. Consequently, the corresponding stakeholders (industry owners, government authorities, factory workers, etc.) can introduce regulatory frameworks and control the proliferation of visual pollution. The open-source images obtained by web crawling, locally collected visual pollutants dataset and implementation code of this work are available at: https://github.com/SadiaAfrin163/Textile-Visual-Pollutants-Dataset. [ABSTRACT FROM AUTHOR]

Many specific problems within the domain of high throughput phenotyping require the accurate localization of plant organs. To track and count plant organs, we propose GCNNMatch++, a Graph Convolutional Neural Network (GCNN) that is capable of online tracking objects from videos. Based upon the GCNNMatch tracker with an improved CensNet GNN, our end-to-end tracking approach achieves fast inference. In order to adapt this approach to flower counting, we collected a large, high-quality dataset of cotton flower videos by leveraging our custom-built MARS-X robotic platform. Specifically, our system can count cotton flowers in the field with 80% accuracy, achieving a Higher-Order Tracking Accuracy (HOTA) of 51.09 and outperforming more generic tracking methods. Without any optimization (such as employing TensorRT), our association model runs in 44 ms on a central processing unit (CPU). On appropriate hardware, our model holds promise for achieving real-time counting performance when coupled with a fast detector. Overall, our approach is useful in counting cotton flowers and other relevant plant organs for both breeding programs and yield estimation. [Display omitted] • A new graph-based multi-object tracking method was developed to count cotton flowers. • The method's performance was assessed on image data collected with a ground robot. • The model's potential for phenotyping flowering traits was investigated. [ABSTRACT FROM AUTHOR]

Banana (Musa spp.) is an extensively favored fruit owing to its affordability and considerable nutritional richness. Ensuring the quality of bananas is essential for meeting consumer expectations and international export standards. Post-harvest handling, particularly grading, plays a pivotal role in sorting and classifying bananas. It is crucial for quality assurance, aligning with consumer preferences, gaining market access, differentiating prices based on quality, minimizing waste, optimizing packaging efficiency and enhancing the overall effectiveness of the supply chain. The manual grading is accountable for a considerable range of post-harvest losses. The Convolutional Neural Network (CNN) is widely recognized as a state-of-the-art computer vision technique for classification tasks. In this investigation, a CNN-based deep learning approach is introduced for the ripening classification of the Nendran banana. This study focused on the development and evaluation of a CNN model using a dataset of 4320 images. Pre-existing Deep Learning (DL) models (VGG16, VGG19, InceptionV3, ResNet50 and EfficientNetB0) were employed for comparison, and the developed model achieved 95 % accuracy. A web application titled 'Banana Ripeness Identification App', was developed using the proposed model. Notably, the proposed model outperformed existing DL models, emphasizing its superior classification accuracy. The study concluded that the 9-layer CNN model is highly effective and surpasses established DL architectures, which can serve as a foundation for the advancement of efficient classification methods for bananas based on their ripeness, consequently enhancing post-harvest management. • Developed a new CNN model for banana ripeness classification. • The developed model outperformed all the existing classification models. • The model can be used in hand-held devices for grading banana. • 4320 images collected and made opensource. • A web application developed to showcase the ability of the developed model. [ABSTRACT FROM AUTHOR]

The Transformer network is the main application area for natural language processing. It has gained traction lately and exhibits potential in the field of computer vision. This cutting-edge method has proven to offer a significant impact on image analysis, a crucial area of computer vision. The transformer's outstanding performance in vision computing places it as an alternative to the convolutional neural network for vision tasks. Transformers have taken center stage in the field of natural language processing. Despite the outstanding performance of transformer networks in natural image processing, their implementation in medical image analysis is gradually gaining roots. This study focuses on the transformer application in natural and medical image analysis. The first part of the study provides an overview of the core concepts of the attention mechanism built into transformers for long-range feature extraction. The study again highlights the various transformer architectures proposed for natural and medical image tasks such as segmentation, classification, image registration and diagnosis. Finally, the paper presents limitations identified in proposed transformer networks for natural and medical image processing. It also highlights prospective study opportunities for further research to better the computer vision domain, especially medical image analysis. This study offers knowledge to scholars and researchers studying computer vision applications as they focus on creating innovative transformer network-based solutions. [ABSTRACT FROM AUTHOR]

• The deep learning models were utilized to detect different e-waste components. • Both TensorFlow and PyTorch modules were utilized for object detection. • An independent evaluation of 240 e-waste objects was performed using YOLO 7 model, which show ∼94 % correct prediction for batch size of 16. An innovative approach is needed to boost the economic value of e-waste by improving metal recovery and facilitating the separation of plastics and valuable metal components. Leveraging deep learning and computer vision technology offers a promising solution for automatically categorizing and sorting e-waste components like copper, printed circuit boards (PCB), steel, glass, and aluminum, presenting significant financial and environmental incentives for increased recycling efforts. In this instance, the real-time object identification algorithms YOLO 7 and 5 were used along with TensorFlow version 2.8.0. The e-waste dataset works incredibly well with TensorFlow and YOLO. For the detection of copper, PCBs, and plastic, the F1 values (F1 score is the harmonic mean of precision and recall; precision is the fraction of relevant instances among the retrieved instances whereas recall or sensitivity is the fraction of relevant instances that were retrieved) were as good as 1.0, whereas the score for steel and aluminum was 0.8. The mean average precision (mAP; for a set of queries, it is the mean of the average precision scores for each query) was 0.96 for all classes, with the highest precision for copper (0.99) followed by PCB (0.981). 240 e-waste objects were independently evaluated using the YOLO v7 model, achieving a remarkable ∼94 % prediction accuracy with a batch size of 16, ensuring robust performance. The real-time e-waste component detection was also done using video clips as well as webcam streaming. Deploying real-time e-waste object detection and sorting can significantly narrow the gap between e-waste accumulation and recycling rates, leading to reduced environmental strain and impact. [Display omitted] [ABSTRACT FROM AUTHOR]

Recently, the successful applications of convolutional neural network (CNN) in computer vision have attracted considerable attentions. Particularly, deep learning models with attention mechanisms have shown impressive performance in hyperspectral (HS) pansharpening. However, most of these existing models follow an early/late-fusion strategy and do not take full advantage of the hierarchical features. In this paper, we specifically design a novel end-to-end multi-level feature fusion network with cross-layer guided attention for HS pansharpening, termed as HP-MFFN, which allows the network to extract the hierarchical features level by level. Specifically, as the different levels of the network have different receptive fields and contain different details, the hierarchical features extracted from the HS image and panchromatic (PAN) image are refined by the cross-layer guided attention fusion module (called CLGAF) to yield more effective spatial-spectral features with fine details and rich semantics. The experimental results conducted on widely-used datasets demonstrate that HP-MFFN provides high-quality pansharpened HS images in terms of perceptually and quantitatively. [ABSTRACT FROM AUTHOR]

Object detection is one significant field of computer vision. The imbalance problem exerts negative effects on achieving satisfactory performance. We reveal two sources of imbalance in existing object detection methods. Correspondingly, we propose our methods in terms of the model architecture and optimization target. Different from general object detection benchmarks, the location distribution of objects with different sizes is unbalanced in many practical applications. In addition, the representation information of different categories of objects is unbalanced. In this paper, we propose a location scale equilibrium module to utilize the prior location scale information and generate more balanced feature maps. More appropriate feature maps are selected and merged for different locations. After merging, feature maps become more consistent in terms of representation content, exerting positive effects on the following classification and regression tasks. For the imbalance caused by similar objects, we propose the repulsive loss to strengthen the punishment. Our method will not treat all categories of objects equally since we take the imbalance between them into consideration. With the enhanced supervision, the training will pay more attention to similar objects. Our proposed model is evaluated on the VisDrone benchmark and UAVDT benchmark. Sufficient experiments are conducted. Our model achieves the highest precision on most evaluation metrics, outperforming the other strong models. [ABSTRACT FROM AUTHOR]

In computer vision, edge and object contour detection is essential for higher-level vision tasks, such as shape matching, visual salience, image segmentation, and object recognition. It has attracted much attention during the past several decades, and many excellent methods have been proposed. In this paper, we make a comprehensive introduction to representative edge and object contour detection methods in the past two decades. Based on the development of these methods, we mainly classify them into two categories: traditional methods and learning-based methods. We further divide traditional methods into local pattern methods, edge grouping methods, active contour models, and bio-inspired methods. Further, we divide learning-based methods into classical learning-based methods and deep learning-based methods. At the same time, we introduce the most popular benchmarks and evaluation measures and quantitatively compare the performances of these promising methods. Moreover, we discuss current challenges in edge and object contour detection and suggest some future trends to bridge gaps with human vision. We believe that this overview will benefit newcomers and promote the development of edge and object contour detection. [ABSTRACT FROM AUTHOR]

The rise of deep learning in today's applications entailed an increasing need in explaining the model's decisions beyond prediction performances in order to foster trust and accountability. Recently, the field of explainable AI (XAI) has developed methods that provide such explanations for already trained neural networks. In computer vision tasks such explanations, termed heatmaps , visualize the contributions of individual pixels to the prediction. So far XAI methods along with their heatmaps were mainly validated qualitatively via human-based assessment, or evaluated through auxiliary proxy tasks such as pixel perturbation, weak object localization or randomization tests. Due to the lack of an objective and commonly accepted quality measure for heatmaps, it was debatable which XAI method performs best and whether explanations can be trusted at all. In the present work, we tackle the problem by proposing a ground truth based evaluation framework for XAI methods based on the CLEVR visual question answering task. Our framework provides a (1) selective, (2) controlled and (3) realistic testbed for the evaluation of neural network explanations. We compare ten different explanation methods, resulting in new insights about the quality and properties of XAI methods, sometimes contradicting with conclusions from previous comparative studies. The CLEVR-XAI dataset and the benchmarking code can be found at https://github.com/ahmedmagdiosman/clevr-xai. • First objective, ground truth-based evaluation framework for XAI methods. • Systematic comparison of 10 popular XAI methods using novel evaluation metrics. • New findings regarding strengths and limitations of particular XAI methods. • Evaluation framework, i.e., code and dataset, are publicly available. [ABSTRACT FROM AUTHOR]

The development of stereo matching algorithm is still one of the challenging problems, especially in ill-posed regions. Hence, this article presents a survey on the algorithm frameworks related to the stereo matching algorithm. Based on the early survey that had been conducted, two major frameworks available in current stereo matching algorithm development, they are traditional and artificial intelligence (AI) frameworks. Most of the traditional methods are very low accuracy compared to the AI-based approach. This can be observed in the standard benchmarking dataset, such as from the KITTI and the Middlebury, where AI methods rank at the top of the accuracy list. Additionally, the trend for solving computer vision problems uses AI or machine learning tools that become more apparent in recent years. Thus, this paper is focusing on the survey between the deep learning frameworks, which is one of the machine learning tools related to the convolutional neural network (CNN). Several mixed approaches between CNN based method and traditional handcraft method, as well as the end to end CNN method also discussed in this paper. [ABSTRACT FROM AUTHOR]

Fine-grained visual categorization in zero-shot setting is a challenging problem in the computer vision community. It requires algorithms to accurately identify fine-grained categories that do not appear during the training phase and have high visual similarity to each other. Existing methods usually address this problem by using attribute information as intermediate knowledge, which provides sufficient fine-grained characteristics of categories and can be transferred from seen categories to unseen categories. However, the learning of attribute visual features is not trivial due to the following two reasons: (i) The visual information about attributes of different types may interfere with the visual feature learning of each other. (ii) The visual characteristics of the same attribute may vary in different categories. To solve these issues, we propose a Multi-Group Multi-Stream attribute Attention network (MGMSA), which not only separates the feature learning of attributes of different types, but also isolates the learning of attribute visual features for categories with big differences in attribute appearance. This avoids the interference between uncorrelated attributes and helps to learn category-specific attribute-related visual features. This is beneficial for distinguishing fine-grained categories with subtle visual differences. Extensive experiments on benchmark datasets show that MGMSA achieves state-of-the-art performance on attribute prediction and fine-grained zero-shot learning. [ABSTRACT FROM AUTHOR]

Establishing the relationships among hierarchical visual attributes of objects in the visual world is crucial for human cognition. The classic convolution neural network (CNN) can successfully extract hierarchical features but ignore the relationships among features, resulting in shortcomings compared to humans in areas like interpretability and domain generalization. Recently, algorithms have introduced feature relationships by external prior knowledge and special auxiliary modules, which have been proven to bring multiple improvements in many computer vision tasks. However, prior knowledge is often difficult to obtain, and auxiliary modules bring additional consumption of computing and storage resources, which limits the flexibility and practicality of the algorithm. In this paper, we aim to drive the CNN model to learn the relationships among hierarchical deep features without prior knowledge and consumption increasing, while enhancing the fundamental performance of some aspects. Firstly, the task of learning the relationships among hierarchical features in CNN is defined and three key problems related to this task are pointed out, including the quantitative metric of connection intensity, the threshold of useless connections, and the updating strategy of relation graph. Secondly, Relational Embedding Convolution (RE-Conv) layer is proposed for the representation of feature relationships in convolution layer, followed by a scheme called use & disuse strategy which aims to address the three problems of feature relation learning. Finally, the improvements brought by the proposed feature relation learning scheme have been demonstrated through numerous experiments, including interpretability, domain generalization, noise robustness, and inference efficiency. In particular, the proposed scheme outperforms many state-of-the-art methods in the domain generalization community and can be seamlessly integrated with existing methods for further improvement. Meanwhile, it maintains comparable precision to the original CNN model while reducing floating point operations (FLOPs) by approximately 50%. [ABSTRACT FROM AUTHOR]

Head injury is among the most severe consequences of fall. Hence, there is a significant gap in fall detection research due to the lack of a dataset that includes furniture-caused occlusion and head injuries. Addressing this gap, this study uses a human-like doll to gather videos of falls with and without head injuries from two different angles. Moreover, a multi-input network was designed for fall detection based on pixel-level human detection from two cameras. To learn from the variations of spatial characteristics of the extracted silhouettes over time, two Convolutional Long Short-Term Memory (ConvLSTM) heads were integrated followed by lightweight backbones with depth-wise separable convolution. The concatenation of the extracted features from both streams enhances the accuracy of fall detection while addressing the problem of occlusion. Upon detecting a fall, our system precisely locates the head, tracks its trajectory, and identifies the intersection between the trajectory and the furniture based on our multi-stream head injury detection network. Key points, such as ears, nose, and shoulders, are instrumental in our head detection system. The proposed fall detection network achieved an F1 score of 88.37%, highlighting the effectiveness of multi-angle feature extraction. By combining the head trajectory and furniture mask into a unified representation, the detection of head injuries was improved with less computation, resulting in a 97% F1 score. Besides our dataset, the fall detection network achieved state-of-the-art results on both the UP Fall and UR Fall datasets, with an F1 score of 94.12% and 100%, respectively. [ABSTRACT FROM AUTHOR]

Lightweight computer vision models with a better balance of speed and accuracy are critical for intelligent system devices. Wearing a mask correctly can effectively prevent the spread of respiratory diseases. However, most of the deep learning-based face mask detection(FMD) algorithms are too large and have limited accuracy, limiting their deployment in real-world application scenarios. Therefore, this paper proposes a novel object detector, lightweight FMD through You Only Look Once (LFMD-YOLO), which can achieve an excellent balance of precision and speed. First, the C3E(CSP Bottleneck with 3 convolutions-ECA) module and MECAPF(max-pooling ECA pyramid-fast) module based on the effective attention mechanism (ECA) are designed to enrich channel information. Further, a new backbone network combining C3E and MECAPF modules is proposed. Second, the weighted bidirectional feature pyramid network based on the C3E module (E-BiFPN) is designed as the feature fusion neck, making full use of multi-scale features to mine more local information and enhancing the representation of small objects of face masks. Third, the model performance is further enhanced by adding detection heads and improving intersection over union (IoU), etc. Finally, two public databases were benchmarked and compared with sixteen other state-of-the-art detection algorithms. The experimental results show that the proposed LFMD-YOLO achieves higher detection accuracy with mAPs of 68.7% and 60.1%, respectively, while having lower parameters and GFLOPs. In addition, we conducted experiments on the Jetson NVIDIA TX2 to further verify the effectiveness of the LFMD-YOLO. The code is available at https://github.com/ddhappy18/LFMD-YOLO. • A lightweight face detection algorithm based on YOLOv5 (LFMD- YOLO) is proposed. • C3E, MECAPF, and E-BiFPN modules are proposed to enrich channel information. • Proposed LFMD-YOLO-N based on model scaling. • The deployment experiments of edge device TX2 verified the application value. • The proposed detector achieved excellent SAT performance. [ABSTRACT FROM AUTHOR]

In recent years, researchers have made significant strides in computer vision by leveraging transformers, achieving remarkable breakthroughs in low-level vision tasks. The inherent long-range dependency of transformers grants them potent remote modeling capabilities, surpassing those of Convolutional Neural Networks (CNNs) and enabling the extraction of global features and accurate semantic structures. However, it has been observed that single transformer frameworks lack sensitivity to high-frequency information in images, resulting in the generation of blurry reconstructed regions. To address this limitation, this paper proposes a novel two-branch Dual Frequency Feature Fusion Network (DF3Net) for image inpainting based on the hierarchical atrous transformer (HAT). Specifically, the head of the dual-frequency convolution (DFC) module decouples the feature maps into low and high-frequency components. The low-frequency factor goes through the proposed HAT branch, while the high-frequency component is input into the gated convolution module branch, effectively capturing both global structural information and local texture details. Finally, the DFC tail fuses the high and low-frequency features to output the reconstructed image. Moreover, the feature fusion between high and low-frequency branches is performed layer-wise within the network, enabling mutual learning between the two branches interactively, ensuring coherence in image semantic structure and texture details. Experimental evaluations on Places2, Paris StreetView, and CelebA-HQ with different mask ratios demonstrate that the proposed method outperforms state-of-the-art methods in enhancing the structural accuracy of image inpainting. It generates semantically reasonable images with fine texture details. • It proposes a double-branch dual frequency feature fusion network combining transformer and CNNs. • It introduces a hierarchical atrous transformer structure for low-frequency feature extraction. • It develops a dual frequency convolution module by applying the Fourier transformation to the feature maps. • It could achieve competitive results on three popular datasets of different sceneries. [ABSTRACT FROM AUTHOR]

From smart sensors on assembly lines to robots performing complex tasks, the fourth industrial revolution is rapidly transforming manufacturing. The growing prominence of 3D cameras in the industry has led the computer vision community to explore innovative ways of integrating depth and color data to achieve higher precision, essential for ensuring product quality in manufacturing. In this study, we introduce an innovative branched convolutional neural network designed to produce high-speed classification of multimodal images, such as RGB-Depth (RGB-D) images. The fundamental concept underlying the branched approach is the specialization of each branch as a dedicated feature extractor for a single modality, followed by their merge (intermediate fusion) to enable effective classification. Feeding our model is our novel multimodal dataset, named CeramicNet, composed of 8 classes that include RGB, depth, and RGB-D variations to enable extensive experimentation and evaluation of the models which, to the best of our knowledge, has not been previously introduced in the computer vision community. We conducted a series of experiments on the CeramicNet dataset. These experiments aimed at fine-tuning the model, assessing the influence of various depth technologies, exploring individual modalities, examining their collective impact, and performing comprehensive data analysis. Comparing our solution against seven widely used models, we achieved remarkable results, securing the top position with a precision of 99.89, with a lead of over 1% against the nearest competitor. What is more, the proposed solution yields an inference time of 127.6 ms — being nearly three times faster than the second-best performer. • RGB-D Branched CNN for multimodal classification of ceramic pieces. • Our approach achieved 99.89% precision, leading by over 1% against other CNNs. • New RGB-D ceramic dataset with 2.5K images divided into 8 classes called CeramicNet. [Display omitted] [ABSTRACT FROM AUTHOR]

• This study unveils a method to create a substantial PAUT scan database from minimal Full Matrix Capture acquisitions. • This database enables Machine Learning models to generalize effectively, even with limited defect variety. • Integrating geometric context alongside PAUT images significantly boosts detection performance. • The model exhibits robust detection performance, even with noisy images and low CNR, highlighting its potential in real-world applications. Phased array ultrasonic testing (PAUT) requires highly trained and qualified personnel to interpret and analyze images. It takes a solid understanding of wave propagation physics to comprehend the generated images. As such, the inspector's judgment and level of experience have a significant impact on the analysis's outcome. In addition, the procedure is prone to error and laborious. AI had shown to be effective in computer vision in a variety of classification and detection tasks. Regarding PAUT, studies have also demonstrated that machine learning may be able to identify defects with a level of accuracy that is on par or even superior to that of trained and qualified inspectors. Nonetheless, the use of computer vision in PAUT remains very limited. The primary cause of this is the challenge accessing large databases of labelled inspections. In fact, a considerable amount of training data is required for machine learning. While it is easy to access sizeable, labelled databases of MRI scans or photographs for instance, that is not the case in PAUT because inspection results are usually confidential. In this project, a large database was generated using mock-ups commonly used to train and evaluate inspectors. The different defects contained in these mock-ups were used to train a machine learning model. The data was acquired with several different probes centered at different frequencies. Each acquisition was performed using Full Matrix Capture (FMC). The post-processing of the data contained in the FMC allows to compute any sectoral scan from its focal laws. As a result, a comprehensive database composed of hundreds of thousands of sectoral scans was generated from these few FMC acquisitions. The completeness of this database facilitated robust training of a defect detection model for PAUT sectoral scans. The evaluation of the model demonstrated its ability to generalize even to defect types it had never been trained on. Furthermore, the detection performance remained consistent even in high noise conditions where the Contrast-to-Noise Ratio (CNR) was very low. [ABSTRACT FROM AUTHOR]

Date fruit (Phoenix dactylifera L.) is known for its diversity in terms of genotypes that differ in shape, color, texture, taste, and other physical and sensory attributes, in addition to biochemical compounds. This richness generates complex, long, and tedious work at the scale of farm as well as in handling, processing, and storage facilities to prepare standardized, and homogenous final products that deal with consumer preferences and needs. To alleviate this, an increased importance and particular trend were observed recently for the use of artificial intelligence (AI) techniques and computer vision technology since they are considered as accurate, objective, non-destructive, fast, and efficient emerging technologies. In this context, this manuscript provides comprehensive insights into key applications of AI techniques and computer vision technology, using different classification models and algorithms developed on the basis (for example) of neural network classifiers, such as convolutional and artificial neural network, on harvest and postharvest assessment of date fruit. Thus, the importance of date fruit dataset used in the AI domain was discussed with link to all accessible datasets in the field of date palm. In addition, available and relevant AI tools and approaches applied to the postharvest of date fruit from maturity distinguishing until storage through other key handling operations, sorting, grading, etc., were detailed in the five sections of this paper. In terms of conclusion, this review showed that these smart technologies can play a substantial role and offer new opportunities to improve the whole value chain of date palm sector. • Harvesting and post-harvest handling impact quality and storability of dates. • These operations are tedious and subjective leading to quantitative/qualitative losses. • Artificial intelligence (AI) techniques are emergent technologies that can reduce these losses. • AI techniques are more accurate, fast, non-destructive and objective. [ABSTRACT FROM AUTHOR]

• Iron ore composition is vital for steelmaking and improving mineral processing. • Accurate segmentation of quartz particles is difficult. • Deep Learning methods were used to segment quartz particles. • Mask R-CNN was trained with 865 images containing quartz particles. • The F1-Score was over 90%. Iron ore characterization is essential for the mineral industry since it provides relevant information, such as ores' chemical composition and the textural and morphological aspects of particles, required for designating a proper mineral processing route. Reflected Light Optical Microscopy (RLOM) is typically used in this field, as it allows the identification of most mineral phases by their different reflectances in a fast and low-cost procedure. Among those phases, quartz, a gangue mineral, presents a big challenge due to its transparency, as it displays a similar hue to the resin used for sample mounting. Thus, even though specialists can visually identify quartz particles, their recognition by automatic image processing has remained elusive. In the present work, a Deep Learning Instance Segmentation model was trained to identify and segment quartz particles in iron ore optical microscopy images, employing the state-of-the-art Mask R-CNN architecture. The model was trained with 865 images containing 1402 quartz particles, and 216 images, containing 338 quartz particles, were used to evaluate the performance. The performance metrics reached a precision of 95.22%, a recall of 88.46%, and an F1-Score of 91.72%. [ABSTRACT FROM AUTHOR]

Automating identification of benthic habitats from imagery, with Machine Learning (ML), is necessary to contribute efficiently and effectively to marine spatial planning. A promising method is to adapt pre-trained general convolutional neural networks (CNNs) to a new classification task (transfer learning). However, this is often inaccessible to a non-specialist, requiring large investments in computational resources and time (for user comprehension and model training). In this paper, we demonstrate a simpler transfer learning framework for classifying broad deep-sea benthic habitats. Specifically, we take an 'off-the-shelf' CNN (VGG16) and use it to extract features (pixel patterns) from benthic images (without further training). The default outputs of VGG16 are then fed in to a Support Vector Machine (SVM), a classical and simpler method than deep networks. For comparison, we also train the remaining classification layers of VGG16 using stochastic gradient descent. The discriminative power of these approaches is demonstrated on three benthic datasets (574–8353 images) from Norwegian waters; each using a unique imaging platform. Benthic habitats are broadly classified as Soft Substrate (sands, muds), Hard Substrate (gravels, cobbles and boulders) and Reef (Desmophyllum pertusum). We found that the relatively simplicity of the SVM classifier did not compromise performance. Results were competitive with the CNN classifier and consistently high, with test accuracy ranging from 0.87 to 0.95 (average = 0.9 (± 0.04)) across datasets, somewhat increasing with dataset size. Impressively, these results were achieved 2.4–5× faster than CNN training and had significantly less dependency on high-specification hardware. Our suggested approach maximises conceptual and practical simplicity, representing a realistic baseline for novice users when approaching benthic habitat classification. This method has wide potential. It allows automated image grouping to aid annotation or further model selection, as well as screening of old-datasets. It is especially suited to offshore scenarios as it can provide quick, albeit crude, insights into habitat presence, allowing adaptation of sampling protocols in near real-time. • Identifying learned CNN image features using a SVM can swiftly assess benthic habitats both in- and ex-situ. • This approach is competitive with a CNN (VGG16) classifier in terms of performance, speed and simplicity for novice users. • Deep benthic habitats were classified with up to 95% accuracy across three unique datasets. • Training an SVM classifier is significantly less dependent on high-specification hardware. • This work represents a more realistic baseline for novice users with a simple, yet time-consuming, classification problem. [ABSTRACT FROM AUTHOR]

The rapid growth of mobile devices has ushered in an era of different device platforms. Different devices require a consistent user experience, especially with similar graphical user interfaces (GUIs). However, the different code bases of the various operating systems as well as the different GUI layouts and resolutions of the various devices pose a challenge for automated software testing. Crowdsourced software testing (CST) has emerged as a viable solution where crowdsourced workers perform tests on their own devices and provide detailed bug reports. Although CST is cost-effective, it is not very efficient and requires a large number of workers for manual testing. The potential of optimizing CST reproduction testing through computer vision remains largely untapped, especially when considering the uniformity of GUI elements on different devices. In this study, we present a novel deep learning model specifically designed to re-identify GUI elements in CST reproduction test scenarios, regardless of the underlying code changes on different devices. The model features a robust backbone network for feature extraction, an innovative attention mechanism with learnable factors to enhance the features of GUI elements and minimize interference from their backgrounds, and a classifier to determine matching labels for these elements. Our approach was validated on a large GUI element dataset containing 31,098 element images for training, 115,704 element images from real apps for testing, and 67 different background images. The results of our experiments underline the excellent accuracy of the model and the importance of each component. This work is a major step forward in improving the efficiency of reproduction testing in CST. The innovative solutions we propose could further reduce labor costs for CST platforms. • ERINet: A novel CNN model for re-identifying GUI elements across devices. • Attention mechanism with Learnable Factor enhances GUI element re-identification. • Dataset: 31,098 training, 115,704 testing GUI images, and 67 backgrounds. • Potential to revolutionize cross-device testing in crowd-sourced environments. [ABSTRACT FROM AUTHOR]

Crowd counting and crowd density estimation methods are of great significance in the field of public security. Estimating crowd density and counting from single image or video frame has become an essential part of a computer vision system in various scenarios. In this paper, we comprehensively review the recent research advancement on crowd counting and density estimation. First of all, we introduce the background of crowd counting and crowd density estimation. Second, the traditional crowd counting methods are summarized. Third, we focus on reviewing the crowd counting and crowd density methods based on convolutional neural network (CNN) models. Next, we report and discuss the experimental results of a number of typical methods on benchmark datasets. Finally, we present the promising future directions of crowd counting and crowd density. [ABSTRACT FROM AUTHOR]

• A novel convolutional neural network is proposed for salient object detection. • Employing guiding module enhances the availability of contextual information. • Experimental results indicate that the proposed module achieves better metric performance. Salient object detection is a fundamental computer vision task. Deep learning techniques have been successfully applied for salient object detection, however, the saliency maps still suffer from incomplete predictions owing to the difficulty and complexity of objects. We propose a novel convolutional neural network based on aggregating contextual information that can improve the accuracy of salient object detection. In order to obtain high-level global information for fine-level feature maps, we add a pyramid pooling module to the proposed network. Moreover, building a guiding module that consists of four context-information feature modules, the algorithm can gain more effective information and enrich the details of saliency maps. Extensive experiments conducted using six benchmark datasets demonstrate that the proposed method outperforms existing state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

Vehicle logo recognition (VLR) is essential in intelligent transportation systems. Although many VLR algorithms have been proposed, efficient and accurate VLR remains challenging in machine vision. Many VLR algorithms explicitly detect the coarse region of the vehicle logo either by offsetting the detected location of the license plate or by training on numerous images with manual bounding-box annotations. However, the results of license plate detection can significantly influence the VLR accuracy, whereas bounding-box annotations are considerably labor-intensive. Thus, we propose a novel category-consistent deep network learning framework for accurate VLR. A convolutional-neural-network-based vehicle logo feature extraction model is proposed to extract deep features by considering both high- and low-level features in an image. Moreover, a novel category-consistent mask learning module is proposed to help the framework to focus on category-consistent regions without relying on license plate detection or manual box annotations. The deep network is trained and optimized iteratively with the objective function incorporating classification loss and category-consistency loss. Extensive experimental evaluations and comparisons on the publicly available HFUT, XMU, CompCars, and VLD-45 datasets demonstrate the feasibility and superiority of the proposed algorithm. [ABSTRACT FROM AUTHOR]

[Display omitted] • In this paper, we present a RDRF-Net for depth estimation with high accuracy. • It introduces a dynamic adaptive selection method of receptive fields for higher accuracy in depth estimation. • It introduces a residual architecture of 3 × 3 kernels instead of parallel 3 × 3 and 5 × 5 kernels to simulate two receptive fields, which leads to model compression and less network parameters. • It introduces a bottleneck architecture to further compress the network without explicit accuracy loss. Image depth estimation is a challenging problem in computer vision, especially considering both high accuracy and low run time. To save run time and maintain high accuracy, we present a new light-weight model in this paper, i.e., a Residual-based Dynamic Receptive Field Network (RDRF-Net). This model can automatically select the receptive fields suitable for different image scales to generate the depth maps with higher fitting degrees. Residual design and bottleneck layers are used to compress the network for reducing run time. Three groups of experiments are performed on the KITTI dataset to test the accuracy, computation time, and the impact of dynamic receptive fields. Experimental results show that RDRF-Net has comparable accuracy with Godard's model and significantly outperforms it in terms of run time. In addition, it performs closely to Pyd-Net in terms of run time and beats Pyd-Net's accuracy. Experiments also demonstrate the beneficial impact of dynamic receptive fields on improving depth estimation accuracy. [ABSTRACT FROM AUTHOR]

The development of techniques and methods for rapidly and reliably detecting and analysing food quality and safety products is of significance for the food industry. Traditional machine learning algorithms based on handcrafted features normally have poor performance due to their limited representation capacity for complex food characteristics. Recently, the convolutional neural network (CNN) emerges as an effective and potential tool for feature extraction, which is considered the most popular architecture of deep learning and has been increasingly applied for the detection and analysis of complex food matrices. In the current review, the structure of CNN, the method of feature extraction based on 1-D, 2-D and 3-D CNN models, and multi-feature aggregation methods are introduced. Applications of CNN as a depth feature extractor for detecting and analyzing complex food matrices are discussed, including meat and aquatic products, cereals and cereal products, fruits and vegetables, and others. In addition, data sources, model architecture and overall performance of CNN with other existing methods are compared, and trends of future studies on applying CNN for food detection and analysis are also highlighted. CNN combined with nondestructive detection techniques and computer vision system show great potential for effectively and efficiently detecting and analysing complex food matrices, and the features based on CNN show better performance and outperform the features handcrafted or those extracted by machine learning algorithms. Although there still remains some challenges in using CNN, it is expected that CNN models will be deployed on mobile devices for real-time detection and analysis of food matrices in future. • The principle and architectures of convolutional neural network are presented. • Feature extraction methods based on convolutional neural network are investigated. • Recent applications of convolutional neural network in food are summarized. • Challenges and future work of convolutional neural network are discussed. [ABSTRACT FROM AUTHOR]

• The proposed network contains an attentive recurrent network and a U-recurrent encoder-decoder network. • The attentive recurrent network generates a haze attention map for highlighting the haze regions. • The U-recurrent encoder-decoder network predicts the clear image with the guidance of the haze attention map. • A new residual pyramid pooling module enlarges the receptive field of the whole network. Haze removal is an important pre-processing step in many computer vision tasks. Convolutional neural networks, especially the U-shaped networks, have shown to be effective in image dehazing. Nevertheless, these networks have three main limitations. First, the relevant haze information, e.g. concentration of haze, is totally ignored. Second, spatial inconsistency and information dilution usually occur when the networks refine the dehazed results with a coarse-to-fine strategy. Third, the receptive field of the network is not large enough to capture structural information. Motivated by these problems, a new attentive U-recurrent encoder-decoder dehazing network is presented, which consists of an attentive recurrent network and a U-recurrent encoder-decoder network. By assuming that haze layers with different depths can be detected by multiple stages, we use an attentive recurrent network to generate the haze attention map for guiding the U-recurrent encoder-decoder network with the concentration of haze to better estimate the clear image. Meanwhile, the features for dehazing are further enhanced and the dehazing results are refined in the U-recurrent encoder-decoder network. This design not only enables spatial consistency but also reduces information dilution with short recurrent pathways. Furthermore, a novel residual pyramid pooling module is also proposed and used in the U-recurrent encoder-decoder network, which provides the network with structural information and with an enlarged receptive field. The experimental results demonstrate that our method outperforms state-of-the-art dehazing algorithms on both synthetic and real hazy images. [ABSTRACT FROM AUTHOR]

Technical skill acquisition is an essential component of neurosurgical training. Educational theory suggests that optimal learning and improvement in performance depends on the provision of objective feedback. Therefore, the aim of this study was to develop a vision-based framework based on a novel representation of surgical tool motion and interactions capable of automated and objective assessment of microsurgical skill. Videos were obtained from 1 expert, 6 intermediate, and 12 novice surgeons performing arachnoid dissection in a validated clinical model using a standard operating microscope. A mask region convolutional neural network framework was used to segment the tools present within the operative field in a recorded video frame. Tool motion analysis was achieved using novel triangulation metrics. Performance of the framework in classifying skill levels was evaluated using the area under the curve and accuracy. Objective measures of classifying the surgeons' skill level were also compared using the Mann–Whitney U test, and a value of P < 0.05 was considered statistically significant. The area under the curve was 0.977 and the accuracy was 84.21%. A number of differences were found, which included experts having a lower median dissector velocity (P = 0.0004; 190.38 ms–1 vs. 116.38 ms–1), and a smaller inter-tool tip distance (median 46.78 vs. 75.92; P = 0.0002) compared with novices. Automated and objective analysis of microsurgery is feasible using a mask region convolutional neural network, and a novel tool motion and interaction representation. This may support technical skills training and assessment in neurosurgery. [ABSTRACT FROM AUTHOR]

Nutrition plays a vital role in maintaining human health. Traditional methods used for assessing food composition & nutritional content often require destructive sample preparation, which can be time-consuming and costly. Therefore, computer vision-based approaches have emerged as promising alternatives that enable rapid and non-destructive analysis of various nutritional parameters in foods. In this review, we summarized computer vision applications in meat processing, grains, fruits and vegetables, and seafood. We reviewed recent advancements in computer vision and deep learning-based algorithms employed for food recognition and nutrient estimation. Various existing food recognition and nutrient estimation datasets are also reviewed. Conventional methods offer some limitations, while vision-based technologies provide quick and non-destructive analysis of food composition & nutritional content. Computer vision and deep neural network architectures provide remarkable accuracy for food nutrient measurement. In conclusion, deep learning-based models are paving the way for a promising future in nutritional and health optimization research. In the future, vision-based technologies are expected to transform food classification and detection by enabling more rapid, affordable, and accurate nutritional analyses. Therefore, computer vision is developing into a useful tool for fast and precise evaluation of food nutrients without enabling samples to be damaged. [Display omitted] • Integrating nutrition analysis with computer vision for non-destructive food quality assessment are reviewed. • Transformer-based and CNN-based hybrid approaches are the accurate nutrient recognition methods in the food industry. • Vision-based technologies have revolutionized food analysis by facilitating faster, cost-effective, and precise methods. [ABSTRACT FROM AUTHOR]

Edge detection is of critical importance for middle-level and high-level tasks in computer vision. Existing edge detection methods usually use VGG16 as the encoding network and achieve exceptional performance through transfer learning, which has the characteristics of high parameters and high computational cost. Researchers have implemented edge detection by designing decoding networks. Unlike existing methods, this paper is inspired by the effective mechanisms of edge detection in the parallel pathway of biological vision and proposes a new lightweight encoding–decoding structure for edge detection networks, which we refer to as XYW-Net. In the encoding network, we drew inspiration from the receptive field properties of X-type cells, Y-type cells, and W-type cells involved in the parallel pathway, and developed more meticulous convolutional models. Based on the structure of the parallel pathway, the three cell modules are combined into a novel encoding network and excellent feature extraction performance is obtained. Within the decoding network, this paper draws inspiration from the feature integration ability of the inferotemporal cortex (IT) and introduces a novel feature integration module to constitute the decoding network. Experiments show that the proposed network in this paper obtains the Optimal Dataset Scale (ODS) = 0.812 on the BSDS500 dataset with only 0.79 M parameters. Additionally, the model has exhibited outstanding ODS performance on various publicly available edge detection datasets, including NYUD, BIPEDv1, and Multicue. Moreover, future research could concentrate on implementing more efficient antagonistic mechanisms and devising networks grounded in higher-level visual cortical mechanisms. This could further augment the performance and efficiency of edge detection networks. We encourage researchers to delve into these directions within the field. The codes are available at https://github.com/PXinTao/XYW-Net. [ABSTRACT FROM AUTHOR]

Surveillance of fish disease is essential to contain the spreading epidemics. Current surveillance mostly relies on molecular biology testing methods, which often requires complex procedures and trained operators. These techniques may hardly to be used in many small aquaculture farms. In this paper, a method based on machine vision combined with a target detection network is proposed. A convolutional neural network (CNN) was developed to detect fish infected with SVCV by analyzing the images. The datasets for the CNN that is implemented from a YOLO v7 deep learning algorithm are information extracted from images containing fish populations. An Auto-MSRCR algorithm was used to adaptively enhance the images to minimize manual intervention. We introduced a novel NAM Attention mechanism integrated with the ELAN module in the original YOLO v7 deep learning algorithm. Both channel attention and spatial attention modules were utilized to suppress the insignificant features in the datasets to achieve a more precise and efficient detection method. Also, a loss function MPDIoU was incorporated to avoid missing detection by this vision-based methodology. After training of the model, the detection testing results show that the NAM-YOLO v7 network achieves over 95% prediction accuracy and 93.8% recall, which is superior than other state-of-art YOLO series models. Also, the time for detection for each image only takes 0.18 s illustrating the integrated modules improved the computation efficiency. This novel technique is of great potential to be applied in small farms for rapid and early detection. It can be a vital supplementary tool in developing more effective surveillance strategy by combined with other established methods. • Deep learning algorithm is used in detecting of fish with SVC symptom • Image enhancement is adopted for the analysis of fish populations • MPDIoU is used to enhance the localization of the fish in the images • The detection performance is improved by using NAM attention mechanism combined with YOLOv7 [ABSTRACT FROM AUTHOR]

Passive design characteristics (PDC) play a pivotal role in reducing the energy burden on households without imposing additional financial constraints on project stakeholders. However, the scarcity of PDC data has posed a challenge in previous studies when assessing their energy-saving impact. To tackle this issue, this research introduces an innovative approach that combines deep learning-powered computer vision with machine learning techniques to examine the relationship between PDC and energy burden in residential buildings. In this study, we employ a convolutional neural network computer vision model to identify and measure key indicators, including window-to-wall ratio (WWR), external shading, and operable window types, using Google Street View images within the Chicago metropolitan area as our case study. Subsequently, we utilize the derived passive design features in conjunction with demographic characteristics to train and compare various machine learning methods. These methods encompass Decision Tree Regression, Random Forest Regression, and Support Vector Regression, culminating in the development of a comprehensive model for energy burden prediction. Our framework achieves a 74.2 % accuracy in forecasting the average energy burden. These results yield invaluable insights for policymakers and urban planners, paving the way toward the realization of smart and sustainable cities. • Employing Convolutional Neural Networks (CNN) to extract passive design features. • Passive design characteristics are associated with average energy burden. • Machine Learning-based methods provide the best energy burden prediction. • Support Vector Regressor (SVR) predicts energy burden with 74.2 % accuracy. • Passive design characteristics are essential for the prediction models. [ABSTRACT FROM AUTHOR]

In recent years, saliency detection is playing vital role in many computer vision tasks. Due to the diversity of visual pattern in natural images, a specific saliency approach is hardly suitable for any type of images. The optimal selection of no-reference saliency map without the pixel-wise ground-truth mask as reference is an important problem, which directly influences the performance of the higher-level tasks. This is essentially a saliency quality assessment problem. Unlike other existing solutions, we propose CascNet for both highly efficient and accurate quality assessment. CascNet is consisted of Applicability Sorting Net (ASNet) and Comparing Net (CompNet), where ASNet serves as an efficient filter to only keep suitable saliency algorithms for the given image with no need of generating saliency maps beforehand, while CompNet is used for further fine assessment with a deep Siamese structure. Comprehensive experiments conducted on the commonly used salient object detection datasets THUS10000, DUT-OMRON and the currently most challenging dataset SOC demonstrate the good performance of the proposed method. [ABSTRACT FROM AUTHOR]

• Conventional quality analysis of emulsions proved subjective and time-consuming. • Machine learning outperformed deep learning predicting in-process emulsion quality. • Droplet characteristics are better predictors compared to pixel intensity. • Random Forest model presented 99.8% accuracy. • The automated approach demonstrated promising potential in reducing over processing. Determining the uniformity and consistency of droplet size in the dispersed phase is key to emulsion stability. Conventional methods, which include manual microscopic evaluation and laser diffraction, have presented many challenges to achieve an accurate evaluation of droplet dispersion in pharmaceutical emulsions. Artificial intelligence techniques have demonstrated potential in overcoming the subjectivity and time-consumption associated with the manual approaches in industry. A new automated machine learning approach is presented in this study to predict in-process emulsion quality from micrographs. Droplet characteristics are extracted from emulsion micrographs using a histogram-based image segmentation technique. Machine learning classification models are developed, with a selected set of droplet characteristics as predictors, via Random Forest, Multinomial Logistic Regression and Vanilla Neural Network to classify the micrographs into four categories. The hyper-parameters of the models are tuned using 10-fold cross validation. A pixel-based Convolutional Neural Network model is also investigated. Random Forest presented the best accuracy of 99.78% compared to the deep learning models, which presented a bias towards the high frequency classes. The automated machine learning approach demonstrated promising potential for inline emulsion quality evaluation. [ABSTRACT FROM AUTHOR]