Search

Precise and automatic plant disease detection represent a fundamental research topic. Indeed, traditional or manual disease detection can be laborious, inaccurate and time-consuming. In this work, a particular attention is given to automatic potato late blight disease detection. In fact, this devastating pathogen leads every year to a significant reduction in potato yields. In addition, since potatoes are a major food source for many nations, decreased production generate a real food insecurity. Therefore, considering these challenges, using advanced techniques in computer vision and machine learning allowed farmers to swiftly and accurately identify disease. The main objective of this research work is to generate a super-resolved labeled dataset (SRD) and evaluate its impact on plant disease detection. Three states of the art object detection methods (Faster-RCNN, Detr and Yolo V8) have been used to conduct an exhaustive evaluation on the effect of using a super-resolved dataset to perform detection. The obtained results show that the detection of potato late blight disease is enhanced using SRD. Training Yolo V8 model on SRD outperform other trained models in detecting very small lesions. In fact, training Yolo V8 on SRD reached higher mAP, lower Loss values and a reasonable inference time, making it suitable for real time applications. [ABSTRACT FROM AUTHOR]

This paper presents Fix-YOLOX (Fixmatch-You Only Look Once X), a semisupervised target detection model that uses a small amount of annotated data for fully supervised training, and adds a semisupervised training module using both pseudolabelling and consistent regularization to prevent overfitting in fully supervised training by using unlabelled data. Additionally, the generalization of the model and its fault tolerance to labelled data are improved. The experimental results show that the proposed semisupervised visual detection algorithm, Fix-YOLOX, can substantially reduce the amount of data annotation required for the target detection task while effectively overcoming the problem caused by annotated data with uneven quality. The YOLOX model achieves 91.95% accuracy with 50% annotated data and an average detection time of 10.4 ms per image/frame, which is consistent with the detection time of original YOLOX. Therefore, the model has good real-time performance and generalizability. [ABSTRACT FROM AUTHOR]

Cross‐dataset object detection methods can adapt to the needs of rapid category expansion in object detection tasks. However, these methods are prone to generate dataset‐aware errors with false alarm objects. This study is aimed to address these issues. A box‐separated multiple‐head module and box‐separated loss function based on the YOLOv8 network are devised to achieve cross‐dataset object detection. Additionally, a sameclass‐aware fusion module to avoid gradient conflicts due to cross‐category conflicts is developed. A multiple‐head fusion module is devised to reduce the number of false alarm objects caused by dataset‐aware errors. A global class‐aware sampler is also designed to adapt to the impact of the imbalanced number of categories and training samples across datasets. The effectiveness of the box‐separated multiple‐head module is verified using cross‐datasets built using the COCO, WiderFace, WiderPerson, and OpenImages V4 datasets. Extensive experiments demonstrate the efficiency and precision of the proposed method. [ABSTRACT FROM AUTHOR]

Traffic sign detection is critical for autonomous driving technology. However, accurately detecting traffic signs in complex traffic environments remains challenge despite the widespread use of one‐stage detection algorithms known for their real‐time processing capabilities. In this paper, the authors propose a traffic sign detection method based on YOLO v8. Specifically, this study introduces the Space‐to‐Depth (SPD) module to address missed detections caused by multi‐scale variations of traffic signs in traffic scenes. The SPD module compresses spatial information into depth channels, expanding the receptive field and enhancing the detection capabilities for objects of varying sizes. Furthermore, to address missed detections caused by complex backgrounds such as trees, this paper employs the Select Kernel attention mechanism. This mechanism enables the model to dynamically adjust its focus and more effectively concentrate on key features. Additionally, considering the uneven distribution of training data, the authors adopted the WIoUv3 loss function, which optimizes loss calculation through a weighted approach, thereby improving the model's detection performance across various sizes and frequencies of instances. The proposed methods were validated on the CCTSDB and TT100K datasets. Experimental results demonstrate that the authors' method achieves substantial improvements of 3.2% and 5.1% on the mAP50 metric compared to YOLOv8s, while maintaining high detection speed, significantly enhancing the overall performance of the detection system. The code for this paper is located at https://github.com/dusongjie/TSD‐YOLO‐Small‐Traffic‐Sign‐Detection‐Based‐on‐Improved‐YOLO‐v8 [ABSTRACT FROM AUTHOR]

In recent years, deep learning methods were good solutions for object detection in synthetic aperture radar (SAR) images. However, the problems of complex scenarios, large object scale differences and imperfect fine‐grained classification in ship detection were intractable. In response, an improved model GDB‐YOLOv5s (Improved YOLOv5s model incorporating global attention mechanism (GAM), DCN‐v2 and BiFusion) is designed. This model introduces deformable convolution networks (DCN‐v2) into the Backbone to enhance the adaptability of the receptive field. It replaces the original Neck's PANet structure with a BiFusion structure to better fuse the extracted multiscale features. Additionally, it integrates GAM into the network to reduce information loss and improve global feature interaction. Experiments were conducted on single‐class dataset SSDD and multi‐class dataset SRSSD‐V1.0. The results show that the GDB‐YOLOv5s model improves mean average precision scores (mAP) significantly, outperforming the original YOLOv5s model and other traditional methods. GDB‐YOLOv5s also improves Precision‐score (P) and Recall‐score (R) for fine‐grained classification to some extent, thereby reducing false alarms and missed detections. It has been proved that the improved model is relatively effective. [ABSTRACT FROM AUTHOR]

Hand gesture recognition is important in human-machine interaction (HMI), enabling interaction with the systems without physically touching them. But current methodologies regarding gesture recognition face different challenges, such as poor lighting conditions, complex backgrounds, lower detection rates, slower speed, etc. The physically handicapped people with fewer fingers find it difficult to engage with desktop applications in case of complex backgrounds. To overcome these concerns, we have chosen the YOLOv5s model, one small version of YOLOv5 (You Only Look Once) object detection algorithm, to detect and classify the hand portion in a real-time scenario. We have fine-tuned the YOLOv5s architecture by freezing some convolutional layers in the backbone portion, which reduces the number of parameters, model size, and inference time. Two datasets, one public (American sign language) and one custom dataset, named 'NITR-Hand gesture' dataset have been utilized to evaluate the suggested work. There is always a trade-off between accuracy and detection speed. Therefore, the accuracy has been slightly reduced after freezing the backbone of YOLOv5s architecture. but the inference time and detection speed have both improved. Experimental results exhibit that our suggested frozen YOLOv5s model has achieved a mean average precision (mAP@50-95) of 92.60%, and the average speed has reached more than 55 frames per second (fps). We have conducted a comparative analysis of our used models with other state-of-the-art methods, and it is noticed that the fine-tuned YOLOv5s have outperformed other models in terms of mAP and inference speed. This prototype is very fruitful for physically impaired persons to interact with the systems in real-time. [ABSTRACT FROM AUTHOR]

In the domain of object detection, the performance of a detector depend heavily on the quality of features extracted by the backbone network. Extraordinary feature representation has significantly improved the performance of the detector. As we all know, proposing a novel backbone structure requires meticulous structural design, a wealth of expert experience and tricks, and consumes a lot of computing resources. Therefore, it is meaningful to effectively leveraging the existing pre-trained backbones and maximize their performance to improve the accuracy of object detection. In this paper, we propose a novel Feature-enhanced Composite Backbone Network to improve the feature representation capability of the backbone, which called FECNet, equipped with Proportional Feature Fusion Module(PFF) and Multi-Granularity Information Aggregation and Interaction Method(MIAM). In particular, FECNet combines the existing backbones, which are connected by PFF, and FECNet pays more attention to extracting the discriminative features related to the object that are suitable for classification and the edge information suitable for bounding box regression through MIAM. Experiments show that FECNet is easily integrated into mainstream detectors and improve their performances. On the COCO 2017 dataset, employing ResNet50 as the foundational backbone, our approach attains a notable 3 percent increment in performance within the Fast R-CNN framework. Simultaneously, our method yields 2.1 percent enhancement when applied to the ResNet101 + Cascade R-CNN. And it's worth noting that applying our approach on the backbone swin transformer which is based on the trasformer structure gets an increase of more than 2 percent. [ABSTRACT FROM AUTHOR]

Introduction: In the field of facility agriculture, the accurate identification of tomatoes at multiple stages has become a significant area of research. However, accurately identifying and localizing tomatoes in complex environments is a formidable challenge. Complex working conditions can impair the performance of conventional detection techniques, underscoring the necessity for more robust methods. Methods: To address this issue, we propose a novel model of YOLOv8-EA for the localization and identification of tomato fruit. The model incorporates a number of significant enhancements. Firstly, the EfficientViT network replaces the original YOLOv8 backbone network, which has the effect of reducing the number of model parameters and improving the capability of the network to extract features. Secondly, some of the convolutions were integrated into the C2f module to create the C2f-Faster module, which facilitates the inference process of the model. Third, the bounding box loss function was modified to SIoU, thereby accelerating model convergence and enhancing detection accuracy. Lastly, the Auxiliary Detection Head (Aux-Head) module was incorporated to augment the network's learning capacity. Result: The accuracy, recall, and average precision of the YOLOv8-EA model on the self-constructed dataset were 91.4%, 88.7%, and 93.9%, respectively, with a detection speed of 163.33 frames/s. In comparison to the baseline YOLOv8n network, the model weight was increased by 2.07 MB, and the accuracy, recall, and average precision were enhanced by 10.9, 11.7, and 7.2 percentage points, respectively. The accuracy, recall, and average precision increased by 10.9, 11.7, and 7.2 percentage points, respectively, while the detection speed increased by 42.1%. The detection precision for unripe, semi-ripe, and ripe tomatoes was 97.1%, 91%, and 93.7%, respectively. On the public dataset, the accuracy, recall, and average precision of YOLOv8-EA are 91%, 89.2%, and 95.1%, respectively, and the detection speed is 1.8 ms, which is 4, 4.21, and 3.9 percentage points higher than the baseline YOLOv8n network. This represents an 18.2% improvement in detection speed, which demonstrates good generalization ability. Discussion: The reliability of YOLOv8-EA in identifying and locating multi-stage tomato fruits in complex environments demonstrates its efficacy in this regard and provides a technical foundation for the development of intelligent tomato picking devices. [ABSTRACT FROM AUTHOR]

• Understanding plastic litter in cities is crucial to preventing marine pollution. • Citizen science is suitable for long-term, extensive data collection. • Despite human error, ∼30 % of data could be used to visualize and quantify litter. • Litter in cities was studied using Pirika and deep learning-based image processing. • Trained models exhibited high precision and recall in classifying litter types. Cities are a major source of litter pollution. Determination of the abundance and composition of plastic litter in cities is imperative for effective pollution management, environmental protection, and sustainable urban development. Therefore, here, a multidisciplinary approach to quantify and classify the abundance of litter in urban environments is proposed. In the present study, litter data collection was integrated via the Pirika smartphone application and conducted image analysis based on deep learning. Pirika was launched in May 2018 and, to date, has collected approximately one million images. Visual classification revealed that the most common types of litter were cans, plastic bags, plastic bottles, cigarette butts, cigarette boxes, and sanitary masks, in that order. The top six categories accounted for approximately 80 % of the total, whereas the top three categories accounted for more than 60 % of the total imaged litter. A deep-learning image processing algorithm was developed to automatically identify the top six litter categories. Both precision and recall derived from the model were higher than 75 %, enabling proper litter categorization. The quantity of litter derived from automated image processing was also plotted on a map using location data acquired concurrently with the images by the smartphone application. Conclusively, this study demonstrates that citizen science supported by smartphone applications and deep learning-based image processing can enable the visualization, quantification, and characterization of street litter in cities. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In the context of autonomous driving, real-time and accurate object detection is crucial for ensuring vehicle safety. To address the problem of severe missed detection and poor accuracy for small targets in complex traffic conditions, a novel object detection algorithm for automatic driving is proposed, named YED-YOLO. Firstly, based on the original algorithm, the efficient multiscale attention (EMA) module is introduced to enhance the attention to the underlying features. Then, the CSPDarknet53 to 2-stage FPN (C2f) module is improved to solve the problem of insufficient receptive fields by adding deformable convolutions (DCNs). Furthermore, a novel intersection over union (IoU) loss function calculation method is proposed, which improves the generalization ability and convergence speed of the model. Finally, the ablation and comparative experiments are conducted using the recall, precision and frames per second (FPS). The experimental results indicate that the algorithm proposed in this paper demonstrates substantial enhancement effet for smaller traffic participants such as pedestrians and cyclists, whose APs increase by 3% and 3.1%, respectively. [ABSTRACT FROM AUTHOR]

How to effectively and efficiently identify multi-scale objects is one of the key challenges in object detection. In order to make the classification and regression of single-stage object detector more accurate, an improved algorithm named you only look once with squeeze-and-excitation, coordinate attention and adaptively spatial feature fusion (SCA-YOLOv4) was proposed based on YOLOv4 algorithm. In this paper, firstly, by studying the different combination models of squeeze-and-excitation (SE) and coordinate attention (CA) module, the optimal detection performance model of SE and CA combination model was determined. On this basis, squeeze-and-excitation coordinate attention (SECA) combination model was embedded in the head module of YOLOv4 algorithm to highlight the useful information of the feature graph and suppress the irrelevant information. Finally, adaptively spatial feature fusion (ASFF) module was embedded between neck and head module to learn how to filter other levels of features in space, so as to retain only useful information in order to correctly integrate different levels of features. The average accuracy of the proposed model was 91.1% and 46.4% on PASCAL VOC2007 and MS COCO2017 datasets respectively, which was 1.9% and 2.9% higher than those of the original YOLOv4 algorithm. Therefore, experiments indicate that the proposed SCA-YOLOv4 model exhibits optimal performance on the PASCAL VOC and COCO datasets. [ABSTRACT FROM AUTHOR]

Most existing Visual Simultaneous Localization and Mapping (VSLAM) systems heavily rely on the assumption of a static world. However, in real-world scenarios, moving objects often reduce the accuracy and robustness of these systems. To mitigate the impact of moving objects on SLAM system performance, this paper introduces GGC-SLAM, a VSLAM system designed for dynamic environments based on the static probability of predictive feature points. Building on the foundation of ORB-SLAM2, this system incorporates a lightweight object detection thread for acquiring semantic information. It also integrates the Grid-based Motion Statistics (GMS) and Graph-Cut RANdom SAmple Consensus (GC-RANSAC) algorithms to enhance the speed and precision of fundamental matrix computation. The system preliminarily predicts the static probability of feature points by combining semantic information with epipolar constraints, then refines the differentiation between dynamic and static feature points using the initial probability information and a Conditional Random Field, ultimately excluding feature points deemed dynamic. We conducted evaluations using the TUM public dataset and in real-world environments, and the results show that GGC-SLAM can effectively handle dynamic feature points in dynamic scenes. While ensuring real-time operation, it demonstrates more accurate localization compared to other advanced dynamic SLAM systems. Particularly in high-dynamic scenes, our system's average absolute trajectory error is reduced by 95.71% compared to ORB-SLAM2. [ABSTRACT FROM AUTHOR]

Given the challenges posed by noise and varying target scales in SAR images, conventional convolutional neural networks often underperform in SAR image detection. To address this, this paper introduces a novel approach. Firstly, a Res-Clo network is proposed for denoising SAR images as a preprocessing step to enhance detection accuracy. Subsequently, an improved network, DML-YOLOv8, is devised based on the YOLOv8 network. The enhancements in the proposed algorithm include several key modifications. Firstly, within the feature extraction layers, a designed MFB module is integrated to effectively broaden the network's receptive field. Next, deformable convolutions are introduced in the feature fusion layers to bolster the network's capability for multi-scale detection. Additionally, a novel loss function, RT-IOU, is designed in feature detection to enhance network inference speed. Finally, a specialized STD small target detection layer is designed to improve detection accuracy for small targets. In practical experiments, it has been shown that the detection method proposed in this paper effectively improves the detection performance of noisy SAR images, and also achieves satisfactory results in multi-scale detection. [ABSTRACT FROM AUTHOR]

Existing methods for fusing visible light and infrared features often focus on separating objects from backgrounds. In unmanned aerial vehicle (UAV) images, background areas typically contain multiple types of small objects, and the dispersed nature of these objects reduces the effectiveness of feature fusion, leading to phenomena such as missed detections and false alarms when dealing with aerial small objects. This paper introduces a bimodal feature fusion network named EBFF-Net, designed to address the issue of UAV image object detection in low-visibility environments. In this study, a shallow learning network is utilized to extract complementary features, and the Parallel Shallow Feature Fusion (PSFF) method is designed to extract and fuse bimodal features. Additionally, a reconfigurable structure with diverse branch blocks is introduced to the bottleneck layer to better capture feature information without increasing the computational burden during inference. Furthermore, leveraging the geometric properties of two-dimensional rectangles and based on an adaptive weighting algorithm, a novel localization loss function is developed. Subjective and objective testing on the VEDAI, DIOR, and VAID datasets validate the efficacy and lightweight. [ABSTRACT FROM AUTHOR]

The image object recognition and detection technology are widely used in many scenarios. In recent years, big data has become increasingly abundant, and big data-driven artificial intelligence models have attracted more and more attention. Evolutionary computation has also provided a powerful driving force for the optimization and improvement of deep learning models. In this paper, we propose an image object detection method based on self-supervised and data-driven learning. Differ from other methods, our approach stands out due to its innovative use of multispectral data fusion and evolutionary computation for model optimization. Specifically, our method uniquely combines visible light images and infrared images to detect and identify image targets. Firstly, we utilize a self-supervised learning method and the AutoEncoder model to perform high-dimensional feature extraction on the two types of images. Secondly, we fuse the extracted features from the visible light and infrared images to detect and identify objects. Thirdly, we introduce a model parameter optimization method using evolutionary learning algorithms to enhance model performance. Validation on public datasets shows that our method achieves comparable or superior performance to existing methods. [ABSTRACT FROM AUTHOR]

Wheat is one of the most common staple crops and every year, a million tons have been exported from India worldwide. Among all states of India, more than 50% of wheat grew in different regions of Punjab. According to the Indian council of agricultural research ("Dare wheat annual report," n.d.), every year more than 10% growth rate of the wheat crop has decreased due to wheat rust diseases (WRD). Thus, the monitoring of the wheat crop diseases is important. This paper presents a panoptic segmentation-based hybrid approach namely as FERSPNET-50 approach for WRD detection in real-time images and a dataset was collected with different weather conditions in various regions of Punjab. With the help of the GNet model, the collected images have been classified into three classes of weather conditions and then this classified weather image has been fed as input to a faster region-based convolutional model (FRCNN). FRCNN model detects the wheat leaves and stems patches. As a part of the FRCNN training process, a semi-automatic annotation method is designed for generating ground truth rust lines. To further improve the accuracy of rust detections over the whole image, a pyramid scene parsing network has been employed to predict the rust diseases at the local level. Additionally, a deep CNN pretrained model has been utilized to determine the orientation of rust segments within each detected patch, and patches that don't support local level patches are viewed as false positives. After the classification of each patch, the amount of rust severity in wheat has been calculated. The experimental results show that the proposed approach has a high precision (0.97) as compared to state of art based YOLOV4(0.88) and RetinaNet (0.82) models for WRD detection. [ABSTRACT FROM AUTHOR]

Enantiospecific effects play an uprising role in chemistry and technical applications. Chiral molecular networks formed by self‐assembly processes at surfaces can be imaged by scanning probe microscopy (SPM). Low contrast and high noise in the topography map often interfere with the automatic image analysis using classical methods. The long SPM image acquisition times restrain Artificial Intelligence‐based methods requiring large training sets, leaving only tedious manual work, inducing human‐dependent errors and biased labeling. By generating realistic looking synthetic images, the acquisition of real datasets is avoided. Two state‐of‐the‐art object detection architectures are trained to localize and classify chiral unit‐cells in a regular molecular chiral network formed by self‐assembly of linear molecular bricks. The comparison of different architectures and datasets demonstrates that the training on purely synthetic data outperforms models trained using augmented datasets. A Faster R‐CNN model trained solely on synthetic data achieved an excellent mean average precision of 99% on real data. Hence this approach and the transfer to real data show high success, also highlighting the high robustness against experimental noise and different zoom levels across the full experimentally reasonable parameter range. The generalizability of this idea is demonstrated by achieving equally high performance on a different structure, too. [ABSTRACT FROM AUTHOR]

In digital agriculture, accurate crop detection is fundamental to developing automated systems for efficient plantation management. For oil palm, the main challenge lies in developing robust models that perform well in different environmental conditions. This study addresses the feasibility of using GAN augmentation methods to improve palm detection models. For this purpose, drone images of young palms (< 5 year-old) from eight different estates were collected, annotated, and used to build a baseline detection model based on DETR. StyleGAN2 was trained on the extracted palms and then used to generate a series of synthetic palms, which were then inserted into tiles representing different environments. CycleGAN networks were trained for bidirectional translation between synthetic and real tiles, subsequently utilized to augment the authenticity of synthetic tiles. Both synthetic and real tiles were used to train the GAN-based detection model. The baseline model achieved precision and recall values of 95.8% and 97.2%. The GAN-based model achieved comparable result, with precision and recall values of 98.5% and 98.6%. In the challenge dataset 1 consisting older palms (> 5 year-old), both models also achieved similar accuracies, with baseline model achieving precision and recall of 93.1% and 99.4%, and GAN-based model achieving 95.7% and 99.4%. As for the challenge dataset 2 consisting of storm affected palms, the baseline model achieved precision of 100% but recall was only 13%. The GAN-based model achieved a significantly better result, with a precision and recall values of 98.7% and 95.3%. This result demonstrates that images generated by GANs have the potential to enhance the accuracies of palm detection models. [ABSTRACT FROM AUTHOR]

Introduction: Efficiently and precisely identifying tomatoes amidst intricate surroundings is essential for advancing the automation of tomato harvesting. Current object detection algorithms are slow and have low recognition accuracy for occluded and small tomatoes. Methods: To enhance the detection of tomatoes in complex environments, a lightweight greenhouse tomato object detection model named S-YOLO is proposed, based on YOLOv8s with several key improvements: (1) A lightweight GSConv&lowbar;SlimNeck structure tailored for YOLOv8s was innovatively constructed, significantly reducing model parameters to optimize the model neck for lightweight model acquisition. (2) An improved version of the α-SimSPPF structure was designed, effectively enhancing the detection accuracy of tomatoes. (3) An enhanced version of the β-SIoU algorithm was proposed to optimize the training process and improve the accuracy of overlapping tomato recognition. (4) The SE attention module is integrated to enable the model to capture more representative greenhouse tomato features, thereby enhancing detection accuracy. Results: Experimental results demonstrate that the enhanced S-YOLO model significantly improves detection accuracy, achieves lightweight model design, and exhibits fast detection speeds. Experimental results demonstrate that the S-YOLO model significantly enhances detection accuracy, achieving 96.60% accuracy, 92.46% average precision (mAP), and a detection speed of 74.05 FPS, which are improvements of 5.25%, 2.1%, and 3.49 FPS respectively over the original model. With model parameters at only 9.11M, the S-YOLO outperforms models such as CenterNet, YOLOv3, YOLOv4, YOLOv5m, YOLOv7, and YOLOv8s, effectively addressing the low recognition accuracy of occluded and small tomatoes. Discussion: The lightweight characteristics of the S-YOLO model make it suitable for the visual system of tomato-picking robots, providing technical support for robot target recognition and harvesting operations in facility environments based on mobile edge computing. [ABSTRACT FROM AUTHOR]

This study analyzes YouTube thumbnails to identify key elements that distinguish different categories and attract viewers, specifically focusing on YouTubers in Thailand. Using a fine-tuned Convolutional Neural Network model named Xception, we classified images into food, IT, and travel categories with 88% accuracy. Object detection models identified visual objects in the thumbnails, and the combined classification and detection results were clustered into three groups using K-means. Analysis of each cluster led to category-specific recommendations for food, IT, and travel images. While our novel method of combining multimodal tabular and image features is applicable to various regions, it requires region-specific training data for optimal performance in object detection. Nonetheless, our work is the first to offer concrete advice to novice creators on generating interest through YouTube thumbnails across different composition aspects compared to successful YouTubers. [ABSTRACT FROM AUTHOR]

Visual defect recognition techniques based on deep learning models are crucial for modern industrial quality inspection. The backbone, serving as the primary feature extraction component of the defect recognition model, has not been thoroughly exploited. High‐performance vision transformer (ViT) is less adopted due to high computational complexity and limitations of computational resources and storage hardware in industrial scenarios. This paper presents LSA‐Former, a lightweight transformer architectural backbone that integrates the benefits of convolution and ViT. LSA‐Former proposes a novel self‐attention with linear computational complexity, enabling it to capture local and global semantic features with fewer parameters. LSA‐Former is pre‐trained on ImageNet‐1K and surpasses state‐of‐the‐art methods. LSA‐Former is employed as the backbone for various detectors, evaluated specifically on the PCB defect detection task. The proposed method reduces at least 18M parameters and exceeds the baseline by more than 2.2 mAP. [ABSTRACT FROM AUTHOR]

In order to address the challenge of the small-scale, small-target, and complex scenes often encountered in offshore remote sensing image datasets, this paper employs an interpolation method to achieve super-resolution-assisted target detection. This approach aligns with the logic of popular GANs and generative diffusion networks in terms of super-resolution but is more lightweight. Additionally, the image count is expanded fivefold by supplementing the dataset with DOTA and data augmentation techniques. Framework-wise, based on the Faster R-CNN model, the combination of a residual backbone network and pyramid balancing structure enables our model to adapt to the characteristics of small-target scenarios. Moreover, the attention mechanism, random anchor re-selection strategy, and the strategy of replacing quantization operations with bilinear interpolation further enhance the model's detection capability at a low cost. Ablation experiments and comparative experiments show that, with a simple backbone, the algorithm in this paper achieves a mAP of 71.2% on the dataset, an improvement in accuracy of about 10% compared to the Faster R-CNN algorithm. [ABSTRACT FROM AUTHOR]

Detecting and tracking personnel onboard is an important measure to prevent ships from being invaded by outsiders and ensure ship security. Ships are characterized by more cabins, numerous equipment, and dense personnel, so there are problems such as unpredictable personnel trajectories, frequent occlusions, and many small targets, which lead to the poor performance of existing multi-target-tracking algorithms on shipboard surveillance videos. This study conducts research in the context of onboard surveillance and proposes a multi-object detection and tracking algorithm for anti-intrusion on ships. First, this study designs the BR-YOLO network to provide high-quality object-detection results for the tracking algorithm. The shallow layers of its backbone network use the BiFormer module to capture dependencies between distant objects and reduce information loss. Second, the improved C2f module is used in the deep layer of BR-YOLO to introduce the RepGhost structure to achieve model lightweighting through reparameterization. Then, the Part OSNet network is proposed, which uses different pooling branches to focus on multi-scale features, including part-level features, thereby obtaining strong Re-ID feature representations and providing richer appearance information for personnel tracking. Finally, by integrating the appearance information for association matching, the tracking trajectory is generated in Tracking-By-Detection mode and validated on the self-constructed shipboard surveillance dataset. The experimental results show that the algorithm in this paper is effective in shipboard surveillance. Compared with the present mainstream algorithms, the MOTA, HOTP, and IDF1 are enhanced by about 10 percentage points, the MOTP is enhanced by about 7 percentage points, and IDs are also significantly reduced, which is of great practical significance for the prevention of intrusion by ship personnel. [ABSTRACT FROM AUTHOR]

In the field of automatic optical inspection (AOI), this study presents innovative strategies to enhance object detection accuracy while minimizing dependence on large annotated datasets. We initially developed a defect detection model using a dataset of 3579 images across 32 categories, created in collaboration with a major Taiwanese panel manufacturer. This model was evaluated using 12,000 ambiguously labeled images, with improvements achieved through data augmentation and annotation refinement. To address the challenges of limited labeled data, we proposed the Adaptive Fused Semi-Supervised Self-Learning (AFSL) method. This approach, designed for anchor-based object detection models, leverages a small set of labeled data alongside a larger pool of unlabeled data to enable continuous model optimization. Key components of AFSL include the Bounding Box Assigner, Adaptive Training Scheduler, and Data Allocator, which together facilitate dynamic threshold adjustments and balanced training, significantly enhancing the model's performance on AOI datasets. The AFSL method improved the mean average precision (mAP) from 43.5% to 57.1% on the COCO dataset and by 2.6% on the AOI dataset, demonstrating its effectiveness in achieving high levels of precision and efficiency in AOI with minimal labeled data. [ABSTRACT FROM AUTHOR]

Vehicle detection in remote sensing images is a crucial aspect of intelligent transportation systems. It plays an essential role in road planning, congestion control, and road construction in cities. However, detecting vehicles in remote sensing images is challenging due to their small size, high density, and noise. Most current detectors that perform well in conventional scenes fail to achieve better results in this context. Thus, we propose a quad-layer decoupled network to improve the algorithm's performance in detecting vehicles in remote sensing scenes. This is achieved by introducing modules such as a Group Focus downsampling structure, a quad-layer decoupled detector, and the GTAA label assignment method. Experiments demonstrate that the designed algorithm achieves a mean average precision (mAP) of 49.4 and operates at a speed of 3.0 ms on the RTX3090 within a multi-class vehicle detection dataset constructed based on the xView dataset. It outperforms various real-time detectors in terms of detection accuracy and speed. [ABSTRACT FROM AUTHOR]

Object detection is applied extensively in various domains, including industrial manufacturing, road traffic management, warehousing and logistics, and healthcare. In ship object detection tasks, detection networks are frequently deployed on devices with limited computational resources, e.g., unmanned surface vessels. This creates a need to balance accuracy with a low parameter count and low computational load. This paper proposes an improved object detection network based on YOLOv5. To reduce the model parameter count and computational load, we utilize an enhanced ShuffleNetV2 network as the backbone. In addition, a split-DLKA module is devised and implemented in the small object detection layer to improve detection accuracy. Finally, we introduce the WIOUv3 loss function to minimize the impact of low-quality samples on the model. Experiments conducted on the SeaShips dataset demonstrate that the proposed method reduces parameters by 71% and computational load by 58% compared to YOLOv5s. In addition, the proposed method increases the mAP@0.5 and mAP@0.5:0.95 values by 3.9% and 3.3%, respectively. Thus, the proposed method exhibits excellent performance in both real-time processing and accuracy. [ABSTRACT FROM AUTHOR]

Timely discovery and disposal of road risk sources constitute the cornerstone of road operation safety. Presently, the detection of road risk sources frequently relies on manual inspections via inspection vehicles, a process that is both inefficient and time-consuming. To tackle this challenge, this paper introduces a novel automated approach for detecting road risk sources, termed the multi-scale lightweight network (MSLN). This method primarily focuses on identifying road surfaces, potholes, and scattered objects. To mitigate the influence of real-world factors such as noise and uneven brightness on test results, pavement images were carefully collected. Initially, the collected images underwent grayscale processing. Subsequently, the median filtering algorithm was employed to filter out noise interference. Furthermore, adaptive histogram equalization techniques were utilized to enhance the visibility of cracks and the road background. Following these preprocessing steps, the MSLN model was deployed for the detection of road risk sources. Addressing the challenges associated with two-stage network models, such as prolonged training and testing times, as well as deployment difficulties, this study adopted the lightweight feature extraction network MobileNetV2. Additionally, transfer learning was incorporated to elevate the model's training efficiency. Moreover, this paper established a mapping relationship model that transitions from the world coordinate system to the pixel coordinate system. This model enables the calculation of risk source dimensions based on detection outcomes. Experimental results reveal that the MSLN model exhibits a notably faster convergence rate. This enhanced convergence not only boosts training speed but also elevates the precision of risk source detection. Furthermore, the proposed mapping relationship coordinate transformation model proves highly effective in determining the scale of risk sources. [ABSTRACT FROM AUTHOR]

In recent years, the safety issues of high-speed railways have remained severe. The intrusion of personnel or obstacles into the perimeter has often occurred in the past, causing derailment or parking, especially in the case of bad weather such as fog, haze, rain, etc. According to previous research, it is difficult for a single sensor to meet the application needs of all scenario, all weather, and all time domains. Due to the complementary advantages of multi-sensor data such as images and point clouds, multi-sensor fusion detection technology for high-speed railway perimeter intrusion is becoming a research hotspot. To the best of our knowledge, there has been no review of research on multi-sensor fusion detection technology for high-speed railway perimeter intrusion. To make up for this deficiency and stimulate future research, this article first analyzes the situation of high-speed railway technical defense measures and summarizes the research status of single sensor detection. Secondly, based on the analysis of typical intrusion scenarios in high-speed railways, we introduce the research status of multi-sensor data fusion detection algorithms and data. Then, we discuss risk assessment of railway safety. Finally, the trends and challenges of multi-sensor fusion detection algorithms in the railway field are discussed. This provides effective theoretical support and technical guidance for high-speed rail perimeter intrusion monitoring. [ABSTRACT FROM AUTHOR]

To address the frequent external damage incidents to transmission line corridors caused by construction machinery such as excavators and cranes, this paper constructs a dataset of external damage hazards in transmission line corridors and proposes a detection method based on YOLO-LSDW for these hazards. Firstly, by incorporating the concept of large separable kernel attention (LSKA), the spatial pyramid pooling layer is improved to enhance the information exchange between different feature levels, effectively reducing background interference on external damage hazard targets. Secondly, in the neck network, the traditional convolution is replaced with a ghost-shuffle convolution (GSConv) method, introducing a lightweight slim-neck feature fusion structure. This improves the extraction capability for small object features by fusing deep semantic information with shallow detail features, while also reducing the model's computational load and parameter count. Then, the original YOLOv8 head is replaced with a dynamic head, which combines scale, spatial, and task attention mechanisms to enhance the model's detection performance. Finally, the wise intersection over union (WIoU) loss function is adopted to optimize the model's convergence speed and detection performance. Evaluated on the self-constructed dataset of external damage hazards in transmission line corridors, the improved algorithm shows significant improvements in key metrics, with mAP@0.5 and mAP@0.5:0.95 increasing by 3.4% and 4.6%, respectively, compared to YOLOv8s. Additionally, the model's computational load and parameter count are reduced, and it maintains a high detection speed of 96.2 frames per second, meeting real-time detection requirements. [ABSTRACT FROM AUTHOR]

Substations play a crucial role in the proper operation of power systems. Online fault diagnosis of substation equipment is critical for improving the safety and intelligence of power systems. Detecting the target equipment from an infrared image of substation equipment constitutes a pivotal step in online fault diagnosis. To address the challenges of missed detection, false detection, and low detection accuracy in the infrared image object detection in substation equipment, this paper proposes an infrared image object detection algorithm for substation equipment based on an improved YOLOv8n. Firstly, the DCNC2f module is built by combining deformable convolution with the C2f module, and the C2f module in the backbone is replaced by the DCNC2f module to enhance the ability of the model to extract relevant equipment features. Subsequently, the multi-scale convolutional attention module is introduced to improve the ability of the model to capture multi-scale information and enhance detection accuracy. The experimental results on the infrared image dataset of the substation equipment demonstrate that the improved YOLOv8n model achieves mAP@0.5 and mAP@0.5:0.95 of 92.7% and 68.5%, respectively, representing a 2.6% and 3.9% improvement over the baseline model. The improved model significantly enhances object detection accuracy and exhibits superior performance in infrared image object detection in substation equipment. [ABSTRACT FROM AUTHOR]

This paper proposes a deep learning-based intelligent detection device for insulation pull rod defects, addressing the issues of low detection accuracy, poor timeliness of intelligent analysis, and the difficulty in preserving detection results. Firstly, by constructing the pull rod defects dataset and training the YOLOv5s network, along with commonly used object detection algorithms in industrial defect detection, the feasibility of deep learning networks for insulation pull rod defects detection is explored. Secondly, the trained model is combined to build an intelligent detection device for pull rod defects, integrating insulation pull rod image acquisition and defect detection into a unified system. The research results demonstrate that the YOLOv5s network can quickly and accurately detect pull rod defects. On the test set constructed in this paper, the detection performance metric mAP@0.5:0.95 of the trained model reached 54.7%. Specifically, the mAP@0.5 score was 86.9% at a threshold of 0.5. The detection speed FPS reached 169.5, significantly improving the detection efficiency and accuracy compared to traditional object detection algorithms. By establishing an organic connection between the image hardware acquisition device and the deep learning network, the existing problems of inefficient detection and difficult storage of detection results in pull rod defects detection methods are effectively addressed. This research provides new insights for detecting insulation pull rod defects. [ABSTRACT FROM AUTHOR]

The technology for object detection in remote sensing images finds extensive applications in production and people's lives, and improving the accuracy of image detection is a pressing need. With that goal, this paper proposes a range of improvements, rooted in the widely used YOLOv7 algorithm, after analyzing the requirements and difficulties in the detection of remote sensing images. Specifically, we strategically remove some standard convolution and pooling modules from the bottom of the network, adopting stride-free convolution to minimize the loss of information for small objects in the transmission. Simultaneously, we introduce a new, more efficient attention mechanism module for feature extraction, significantly enhancing the network's semantic extraction capabilities. Furthermore, by adding multiple cross-layer connections in the network, we more effectively utilize the feature information of each layer in the backbone network, thereby enhancing the network's overall feature extraction capability. During the training phase, we introduce an auxiliary network to intensify the training of the underlying network and adopt a new activation function and a more efficient loss function to ensure more effective gradient feedback, thereby elevating the network performance. In the experimental results, our improved network achieves impressive mAP scores of 91.2% and 80.8% on the DIOR and DOTA version 1.0 remote sensing datasets, respectively. These represent notable improvements of 4.5% and 7.0% over the original YOLOv7 network, significantly enhancing the efficiency of detecting small objects in particular. [ABSTRACT FROM AUTHOR]

Precision agriculture-based orchard management relies heavily on the accurate delineation of tree canopies, especially for high-value crops like mangoes. Traditional GIS and remote sensing methods, such as Object-Based Imagery Analysis (OBIA), often face challenges due to overlapping canopies, complex tree structures, and varied light conditions. This study aims to enhance the accuracy of mango orchard mapping by developing a novel deep-learning approach that combines fine-tuned object detection and segmentation techniques. UAV imagery was collected over a 65-acre mango orchard in Multan, Pakistan, and processed into an RGB orthomosaic with a 3 cm ground sampling distance. The You Only Look Once (YOLOv7) framework was trained on an annotated dataset to detect individual mango trees. The resultant bounding boxes were used as prompts for the segment anything model (SAM) for precise delineation of canopy boundaries. Validation against ground truth data of 175 manually digitized trees showed a strong correlation ( R 2 = 0.97), indicating high accuracy and minimal bias. The proposed method achieved a mean absolute percentage error (MAPE) of 4.94% and root mean square error (RMSE) of 80.23 sq ft against manually digitized tree canopies with an average size of 1290.14 sq ft. The proposed approach effectively addresses common issues such as inaccurate bounding boxes and over- or under-segmentation of tree canopies. The enhanced accuracy can substantially assist in various downstream tasks such as tree location mapping, canopy volume estimation, health monitoring, and crop yield estimation. [ABSTRACT FROM AUTHOR]

In recent years, research on adversarial attack techniques for remote sensing object detection (RSOD) has made great progress. Still, most of the research nowadays is on end-to-end attacks, which mainly design adversarial perturbations based on the prediction information of the object detectors (ODs) to achieve the attack. These methods do not discover the common vulnerabilities of the ODs and, thus, the transferability is weak. Based on this, this paper proposes a foreground feature approximation (FFA) method to generate adversarial examples (AEs) that discover the common vulnerabilities of the ODs by changing the feature information carried by the image itself to implement the attack. Specifically, firstly, the high-quality predictions are filtered as attacked objects using the detector, after which a hybrid image without any target is made, and the hybrid foreground is created based on the attacked targets. The images' shallow features are extracted using the backbone network, and the features of the input foreground are approximated towards the hybrid foreground to implement the attack. In contrast, the model predictions are used to assist in realizing the attack. In addition, we have found the effectiveness of FFA for targeted attacks, and replacing the hybrid foreground with the targeted foreground can realize targeted attacks. Extensive experiments are conducted on the remote sensing target detection datasets DOTA and UCAS-AOD with seven rotating target detectors. The results show that the mAP of FFA under the IoU threshold of 0.5 untargeted attack is 3.4% lower than that of the advanced method, and the mAP of FFA under targeted attack is 1.9% lower than that of the advanced process. [ABSTRACT FROM AUTHOR]

In recent years, object detection in unmanned aerial vehicle (UAV) imagery has been a prominent and crucial task, with advancements in drone and remote sensing technologies. However, detecting targets in UAV images pose challenges such as complex background, severe occlusion, dense small targets, and lighting conditions. Despite the notable progress of object detection algorithms based on deep learning, they still struggle with missed detections and false alarms. In this work, we introduce an MCG-RTDETR approach based on the real-time detection transformer (RT-DETR) with dual and deformable convolution modules, a cascaded group attention module, a context-guided feature fusion structure with context-guided downsampling, and a more flexible prediction head for precise object detection in UAV imagery. Experimental outcomes on the VisDrone2019 dataset illustrate that our approach achieves the highest A P of 29.7% and A P 50 of 58.2%, surpassing several cutting-edge algorithms. Visual results further validate the model's robustness and capability in complex environments. [ABSTRACT FROM AUTHOR]

Adversarial attacks involve introducing minimal perturbations into the original input to manipulate deep learning models into making incorrect network predictions. Despite substantial interest, there remains insufficient research investigating the impact of adversarial attacks in real-world scenarios. Moreover, adversarial attacks have been extensively examined within the digital domain, but adapting them to realistic scenarios brings new challenges and opportunities. Existing physical world adversarial attacks often look perceptible and attention-grabbing, failing to imitate real-world scenarios credibly when tested on object detectors. This research attempts to craft a physical world adversarial attack that deceives object recognition systems and human observers to address the mentioned issues. The devised attacking approach tried to simulate the realistic appearance of stains left by rain particles on traffic signs, making the adversarial examples blend seamlessly into their environment. This work proposed a region reflection algorithm to localize the optimal perturbation points that reflected the trusted regions by employing the trust region optimization with a multi-quadratic function. The experimental evaluation reveals that the proposed work achieved an average attack success rate (ASR) of 94.18%. Experimentation underscores its applicability in a dynamic range of real-world settings through experiments involving distance and angle variations in physical world settings. However, the performance evaluation across various detection models reveals its generalizable and transferable nature. The outcomes of this study help to understand the vulnerabilities of object detectors and inspire AI (artificial intelligence) researchers to develop more robust and resilient defensive mechanisms. [ABSTRACT FROM AUTHOR]

As video surveillance systems become increasingly essential for railway operations, accurate and precise performance testing is crucial. Traditional methods for response latency testing rely on manual readings with millisecond-level clocks, which can lead to compatibility issues, software crashes, and potential security risks. To address these challenges, this paper proposes a response latency testing method based on object detection for railway video surveillance systems. The response latency test method includes two application scenarios: real-time video call and pan–tilt–zoom camera control response. By leveraging the YOLO-V5 model and object detection techniques, the response speed of railway video surveillance systems is effectively evaluated, ensuring testing precision. Experimental results validate the efficiency and feasibility of the proposed approach, emphasizing its enhanced stability and compatibility compared to traditional methods. The proposed approach offers an innovative solution for testing the response lantency of railway video surveillance systems, contributing to the enhancement and optimization of railway operations. [ABSTRACT FROM AUTHOR]

Crack is one of the main factors threatening the safety of the dam. Automatic image object detection is the main way of underwater dam crack detection. However, the traditional methods have problems with low crack detection speed, high false alarm rate, and poor robustness. In addition, the existing methods cannot get a satsifying detection result with a high detection speed. To solve these problems, we propose an efficient dam crack detection method for underwater scenes, called CrackYOLO. Firstly, to better integrate the multi-scale features without incurring excessive computational costs, we propose a feature fusion module in CrackYOLO. Next, we re-design the skip-connection in the network to get better features, compressing the overall model parameters. Then, we propose a feature extraction module called Res2C3, which combines semantic and location information. After that, we proposed a BCAtt to make features focus on both channel and location information. Finally, according to the characteristics of dam underwater crack images, we use a genetic algorithm to select the best value of hyperparameters of the model. The experimental results show that the proposed method detects underwater dam cracks robustly with less computational cost. Our CrackYOLO can get 94.3% mAP[0.5] and 151 FPS in underwater crack detection task which can achieve a real-time detection in practice. [ABSTRACT FROM AUTHOR]

Unmanned aerial vehicle (UAV) image object detection is a challenging task, primarily due to various factors such as multi-scale objects, a high proportion of small objects, significant overlap between objects, poor image quality, and complex and dynamic scenes. To address these challenges, several improvements were made to the YOLOv8 model. Firstly, by pruning the feature mapping layers responsible for detecting large objects in the YOLOv8 model, significant reduction in computational resources was achieved, rendering the model more lightweight. Simultaneously, a detection head fused with self-attention was introduced simultaneously to enhance the detection capability for small objects. Secondly, the introduction of space depth convolution in place of the original convolutional striding and pooling operations facilitates more effective preservation of details in low-resolution images and small objects. Lastly, a multi-level feature fusion module was designed to merge feature maps from different network layers, enhancing the network's representation capability. Results on the Visdrone dataset demonstrate that the proposed model achieved a significant 4.7% improvement in mAP50 compared to YOLOv8, while reducing the parameter count to only 39% of the original model. Moreover, transfer experiments on the TT100k dataset showed a 3.2% increase in mAP50, validating the effectiveness of the improved model for small object detection tasks in UAV images. Our code is made available at . [ABSTRACT FROM AUTHOR]

As an essential object detection application, road damage detection aims to identify and mark road damage. Timely maintenance of detected damage can improve road safety. However, the proportions of damage area of the image is very diffident for the variety of the road damage textures and shapes. Additionally it is a challenge to localize the road damage accurately for the blurring of the damaged regions caused by the external environmental factors. In this study, we propose a Road Damage Detector with a Local Sensing Feature Network (LSF-RDD), which constructs a Local Sensing Feature Network (LSF-Net) as a neck to fuse multi-scale features extracted from the backbone network and can focus on the location of the damaged area. First, the CSP-Darknet53 backbone network extracts the feature maps of three scales layer-by-layer from the input images. Second, these three feature maps are input into LSF-Net for multi-scale feature fusion to generate three local feature representations. LSF-Net comprises four interconnected blocks, enabling top-down and bottom-up feature fusion. Feature maps from the backbone perform multi-scale feature fusion through connections between different blocks. Finally, three local feature representations are sent into the detection head for detection. Experiments show that LSF-RDD performs well on the adopted datasets, especially on the China_motorbike dataset of RDD2022, with mAP@0.5 reaching 94.4%. [ABSTRACT FROM AUTHOR]

This study presents a novel Seeds-Counting pipeline harnessing deep learning algorithms to facilitate the automation of yield prediction prior to harvesting, a crucial component of the breeding process. Unlike existing methods that often cater to a single seed species or those with similar shapes, our approach is capable of accurately estimating the number of seeds across a diverse range of species. The pipeline incorporates a classification network for seed image categorization, along with object detection models specifically tailored to accommodate the morphologies of different seeds. By integrating a seed classifier, three distinct seed detectors, and post-processing filters, our method not only showcases exceptional accuracy but also exhibits robust generalization capabilities across various conditions. Demonstrating an error rate of less than 2% in the test set and achieving accuracy rates exceeding 97% in the extended set, the proposed pipeline offers a viable and efficient solution for high-throughput phenotyping and precision agriculture, effectively overcoming the challenges posed by the diverse morphologies of seeds. [ABSTRACT FROM AUTHOR]

For channel and spatial feature map C×W×H in object detection task, its information fusion usually relies on attention mechanism, that is, all C channels and the entire space W×H are all compressed respectively via average/max pooling, and then their attention weight masks are obtained based on correlation calculation. This coarse-grained global operation ignores the differences among multiple channels and diverse spatial regions, resulting in inaccurate attention weights. In addition, how to mine the contextual information in the space W×H is also a challenge for object recognition and localization. To this end, we propose a Fine-Grained Dual Level Attention Mechanism joint Spacial Context Information Fusion module for object detection (FGDLAM&SCIF). It is a cascaded structure, firstly, we subdivide the feature space W×H into n (optimized as n = 4 in experiments) subspaces and construct a global adaptive pooling and one-dimensional convolution algorithm to effectively extract the feature channel weights on each subspace respectively. Secondly, the C feature channels are divided into n (n = 4) sub-channels, and then a multi-scale module is constructed in the feature space W×H to mine context information. Finally, row and column coding is used to fuse them orthogonally to obtain enhanced features. This module is embeddable, which can be transplanted into any object detection network, such as YOLOv4/v5, PPYOLOE, YOLOX and MobileNet, ResNet as well. Experiments are conducted on the MS COCO 2017 and Pascal VOC 2007 datasets to verify its effectiveness and good portability. [ABSTRACT FROM AUTHOR]

A large number of objects of different scales appear aggregative distribution in remote sensing images, which brings two challenges to remote sensing object detection. The first challenge arises from the presence of objects with large scale differences within the same scene, which complicates the detector's ability to balance the detection performance between large and small objects. The second challenge is that the distribution of dense objects on remote sensing images with complex backgrounds can reduce the feature representation capability of the network, leading to false detections and missed detections of objects. To address the two challenges present in remote sensing object detection, we propose a new object detection network DA-GLANet (Direction-Aware Multi-Branch Attention and Gaussian Label Assignment Balance Net). For the first challenge, we designed a multi-scale feature fusion module to fully utilize semantic information at different scales, thereby improving the detection accuracy of multi-scale objects. Subsequently, to tackle the second challenge, we proposed a multi-branch attention module with direction-aware capabilities. This module can enhance the network's feature representation of objects of interest by generating attention maps that are both direction-aware and position-sensitive. In addition, we propose a label assignment strategy based on Gaussian distribution to address the imbalance of positive and negative labels caused by the complexity of remote sensing images. These methods work synergistically to improve aggregative object detection accuracy. We have conducted extensive experiments on the DOTA dataset, Levir dataset, and WHU buildings change detection dataset to demonstrate the superiority of our method. On the DOTA dataset, our method improves mAP by 0.4%, 0.3%, and 0.2% compared to the top three methods. On the Levir dataset, our method improves mAP by 2.5%, 2.2%, and 1.6% compared to the top three methods. [ABSTRACT FROM AUTHOR]

The use of general target detection algorithms for small-target detection is computationally costly and has a high missed detection rate. A lightweight small-target detection model based on YOLOv5 is proposed to address this issue.First, a maximum pooling layer is introduced to reduce the number of calculations. Second, Shuffle_Conv is designed to replace the ordinary convolutional layer to reduce model parameters. To further compress the model, the Add fusion method is used in the C3 module, while the GAC3 layer is designed with GhostNet. Finally, Mosaic_9 is introduced to improve the small-target detection without increasing the number of calculations. Compared with YOLOv5, computation and parameters of the improved model are reduced by 84.9% and 39.1%, respectively, and the accuracy is improved by 2%, which is more obvious than that of the original model. [ABSTRACT FROM AUTHOR]

Due to the relatively high shooting altitude of unmanned aerial vehicles (UAV), the captured images often contain a multitude of small‐scale targets. To solve the problems of small target scale, lack of semantic information, and high miss detection in drone target detection, in this paper we proposed a more effective unmanned aerial vehicle small target detection algorithm(MEU‐YOLOv5) based on YOLOv5s. Firstly, an efficient global contextual module is proposed to enhance the algorithm's performance in feature extraction while reducing the excessive loss of shallow features. Secondly, a small‐scale target detector is added to enhance the algorithm's detection capability for smaller targets. Lastly, a recursive multi‐level feature fusion path is introduced to better fuse the shallow and deep features of the images, reducing overfitting and improving the algorithm's generalizability and robustness. Experimental results demonstrated that compared to YOLOv5s, MEU‐YOLOv5 achieves a 7.4% improvement in mAP@0.5 and a 4.9% improvement in mAP@0.5:0.95. Additionally, the overall performance of this algorithm surpassed various algorithms in the YOLO series, including YOLOv3, YOLOv5l, YOLOv5m, and YOLOv8s. © 2024 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

Active learning has achieved great success in image classification because of selecting the most informative samples for data labeling and model training. However, the potential of active learning has been far from being realised in object detection due to its unique challenge in utilizing localization information. A popular compromise is to simply take active classification learning over detected object candidates. To consider the localization information of object detection, current effort usually falls into the model-dependent fashion, which either works on specific detection frameworks or relies on additionally designed modules. In this paper, we propose model-agnostic Object Recognition Consistency in Regression (ORCR), which can holistically measure the uncertainty information of classification and localization of each detected candidate from object detection. The philosophy behind ORCR is to obtain the detection uncertainty by calculating the classification consistency through localization regression at two successive detection scales. In the light of the proposed ORCR, we devise an active learning framework that enables an effortless deployment to any object detection architecture. Experimental results on the PASCAL VOC and MS-COCO benchmarks show that our method achieves better performance while simplifying the active detection process. [ABSTRACT FROM AUTHOR]

Search and rescue (SaR) is challenging, due to the unknown environmental situation after disasters occur. Robotics has become indispensable for precise mapping of the environment and for locating the victims. Combining flying and ground robots more effectively serves this purpose, due to their complementary features in terms of viewpoint and maneuvering. To this end, a novel, cost-effective framework for mapping unknown environments is introduced that leverages You Only Look Once and video streams transmitted by a ground and a flying robot. The integrated mapping approach is for performing three crucial SaR tasks: localizing the victims, i.e., determining their position in the environment and their body pose, tracking the moving victims, and providing a map of the ground elevation that assists both the ground robot and the SaR crew in navigating the SaR environment. In real-life experiments at the CyberZoo of the Delft University of Technology, the framework proved very effective and precise for all these tasks, particularly in occluded and complex environments. [ABSTRACT FROM AUTHOR]