Search

This study addresses the current challenges in obtaining preliminary cost estimates for hair transplant procedures, which often involve time-consuming and frustrating communication between clinics and patients. Our goal is to streamline this process by integrating image processing and computer visualization techniques into an application tailored for individuals seeking hair transplantation services. Specifically, our focus is on Asian men experiencing hair loss in Stages 1 to 3. To achieve this, we have developed an image acquisition app, enhanced images as necessary using GFP-GAN, and performed image segmentation with BiSeNet. Additionally, we utilized MediaPipe to create new hairline projections for individuals with receding hairlines. This comprehensive approach allows us to calculate the hair recipient area and the required number of hair grafts accurately. Comparing our estimations to actual measurements, our method shows errors ranging from -15% to +25%, whereas estimations provided by hair clinics range from -35% to +18%. This highlights the superior accuracy of our individual-specific estimations compared to the general guidelines found on internet sources, enabling a more precise cost estimate for patients. [ABSTRACT FROM AUTHOR]

The surveillance and management of cargo fleets is a crucial objective of intelligent transportation systems. Load, especially overload, has a destructive effect on roads and bridges, and monitoring it can increase the life of road surface and its structure. For low-end hardware with lack of CPU power and no GPU support, this paper presents a rapid method to detect whether heavy vehicles have loads or not; then it proposes a fast method for classifying load types to distinguish soil and construction waste from other miscellaneous loads for heavy weight vehicles. This paper applies a method for classifying cargo types using image processing and texture image classification. This method extracts features for statistical analysis of texture images based on gray-level co-occurrence matrices and local binary patterns. The classification is carried out by support vector machine, k-nearest neighborhood, K-mean, artificial neural networks and random forest classifiers. A large number of positive and negative patterns have been used to train these classifiers. We compare the performance of proposed extracted features and classifiers. The simulation results demonstrate that soil and construction waste can be identified from other miscellaneous loads effective in real-time implementation. [ABSTRACT FROM AUTHOR]

The human body comprises various essential organs, including the lungs, which are responsible for the exchange of gases during respiration. However, the modern lifestyle and environmental pollution have contributed to the deterioration of lung health. To address this issue, The medical field has extensively investigated image processing methods for the identification and examination of lung diseases. In our research, we have employed a specific Convolutional Neural Network (CNN) architecture known as You Look Only Once (YOLO) Version 6 for lung cancer detection. We have further optimized the CNN weights using Particle Swarm Optimization(PSO) with dual adaptation and dual validation. The proposed method involves preprocessing the LUNA 16 Challenge Dataset. The dataset used in this study, known as LUNA-16, comprises CT scans obtained from a cohort of 888 patients. In total, the dataset includes 1,181 individual scans, providing a substantial amount of data for analysis and evaluation. Our objective is to develop a model capable of classifying individuals as either having lung cancer or not. Remarkably, our model achieves an accuracy of 82.79% on the LUNA 16 Dataset, which is significant improvement from the previous used methodologies like ANN with 74% accuracy and YOLOv3 with an accuracy of 80.6% showcasing its effectiveness in predicting lung cancer. [ABSTRACT FROM AUTHOR]

One of the most challenging, interesting, and influential areas in image processing is image enhancement. Image enhancement techniques manipulate the existing image so as to ameliorate the quality as well as the visual appearance of the image to the viewer. Different types of image enhancement methods are utilized to tackle the complex problems of image visualization in medical imaging. Many imaging techniques are available, such as CT scans, magnetic resonance imaging, X-rays, and others. MRI is a kind of scan that uses strong magnetic fields and radio waves to capture images of the internal structure of the patient's body. Medical imaging is an exceptionally normal and fundamental medium for clinical experts to conclude illnesses with respect to unseen regions inside the body. In many situations, these images suffer from low contrast and bad illumination. To overcome these problems of low contrast and poor illumination, this paper presents an enhancement scheme using a modified type-2 fuzzy set for MRI images. The results of the proposed scheme are shown in terms of both qualitative and quantitative analysis. All the experiments are carried out for a fixed value of a parameter β = 0.7 . For qualitative analysis, results are visualized with state-of-the-art methods and for quantitative analysis, PSNR, SSIM, AMBE, REC and PL are used. Qualitative and quantitative analysis bear witness to the fact that the performance of the proposed scheme is better in many places in comparison to other existing methods. [ABSTRACT FROM AUTHOR]

Quantum computers can be used for efficient storage, processing and transmission of digital images. It involves exploitation of quantum properties to represent and process the images in a quantum computer. The first important step to process images in a quantum computer is the representation of images in quantum states. In this paper, we introduce two efficient implementation of algorithms to store the bit-planes of an image in quantum states, one for grayscale images and other for color images. The proposed implementations can be used to improve the effectiveness of image processing algorithms such as color complement operation, edge detection, image compression and so on. The proposed implementation has 10% less quantum cost and four times less time complexity compared to BRQI model for greyscale images, and 30% less quantum cost and four times less time complexity compared to QRCI model for color images. [ABSTRACT FROM AUTHOR]

Nowadays, Convolutional neural networks (CNNs) have achieved relevant results on several data sciences-related tasks, such as image processing and pattern recognition. However, CNNs contain an immense number of parameters which often leads to a huge redundancy, overfitting, and a significant amount of memory. In this paper, we aim to present a multi-objective optimization model for kernels redundancy reduction in convolutional neural networks. In fact, the suggested approach, named MOFM-CNN, allows to minimize redundant feature maps using a set of decision control variables. MOFM-CNN is composed of two objectives where in the first one, the decision variables are technically introduced in the cross-entropy function in order to evaluate the impact of each feature map on the CNNs training. In the second one, the control parameters are used to calculate the proportion of active feature maps, that is related to the complexity of the model. The resultant problem is manipulated and solved using non dominated sorting genetic algorithm (NSGA-II). The performance of our proposal is demonstrated visually and numerically for both classification and features maps optimization. [ABSTRACT FROM AUTHOR]

Visual observation and dermoscopic analysis are the most common methods of diagnosing skin cancer. In advanced stages, melanomas spread faster and are less responsive to treatment. Because different lesions in the skin appear similar to one another and sometimes errors in identification occur, the accuracy of diagnosis will decrease significantly when the amount of received images is large. However, the proposed methods for estimating skin lesions and their separation from melanoma have uncertainties and are not generalizable. This paper proposes an optimal decision tree (DT)-based approach, including fuzzy-ID3-pValue and Bayes learning algorithms, which overcomes these challenges. When classifying images, DTs employ a multi-stage procedure to partition the feature space, which enhances their ease of use, precision, and speed. Inference engines are used in fuzzy logic to derive logical deductions about knowledge, which facilitates learning DT and Bayesian learning. Taking advantage of the DT dependency structure, we present a novel fuzzy DT for extracting precise and collaborative fuzzy rules. Furthermore, to emphasize the cohesive nature of the laws, a weighted method is employed. Besides, the inference engine system is constructed through deductive and inductive inference engines. The proposed method is verified using the ISIC-2019 dataset as well as the PH2 images, both of which contain dermoscopic images of multiple lesions. The proposed method provides efficient results with 96% and 88% accuracy for ISIC-2019 and PH2 data, respectively. [ABSTRACT FROM AUTHOR]

Leukemia is a type of blood tumour that occurs because of abnormal enhancement in WBCs (white blood cells) in the bone marrow of the human body. Blood-forming tissue cancer influences the lymphatic and bone marrow system. The early diagnosis and detection of leukaemia, i.e., the accurate difference of malignant leukocytes with little expense at the beginning of the disease, is a primary challenge in the disease analysis field. Despite the higher occurrence of leukemia, there is a lack of flow cytometry tools, and the procedures accessible at medical diagnostics centres are time-consuming. Distinct researchers have implemented computer-aided diagnostic (CAD) and machine learning (ML) methods for laboratory image analysis, aiming to manage the restrictions of late leukemia analysis. This study proposes a new Falcon optimization algorithm with deep convolutional neural network for Leukemia detection and classification (FOADCNN-LDC) technique. The main objective of the FOADCNN-LDC technique is to classify and recognize leukemia. The FOADCNN-LDC technique utilizes a median filtering (MF) based noise removal process to eradicate the image noise. Besides, the FOADCNN-LDC technique employs the ShuffleNetv2 model for the feature extraction process. Moreover, the detection and classification of the leukemia process are performed by utilizing the convolutional denoising autoencoder (CDAE) model. The FOA is implemented to select the hyperparameter of the CDAE model. The simulation process of the FOADCNN-LDC approach is performed on a benchmark medical dataset. The investigational analysis of the FOADCNN-LDC approach highlighted a superior accuracy value of 99.62% over existing techniques. [ABSTRACT FROM AUTHOR]

Background: Monkeypox is a viral disease caused by the monkeypox virus (MPV). A surge in monkeypox infection has been reported since early May 2022, and the outbreak has been classified as a global health emergency as the situation continues to worsen. Early and accurate detection of the disease is required to control its spread. Machine learning methods offer fast and accurate detection of COVID-19 from chest X-rays, and chest computed tomography (CT) images. Likewise, computer vision techniques can automatically detect monkeypoxes from digital images, videos, and other inputs. Objectives: In this paper, we propose an automated monkeypox detection model as the first step toward controlling its global spread. Materials and method: A new dataset comprising 910 open-source images classified into five categories (healthy, monkeypox, chickenpox, smallpox, and zoster zona) was created. A new deep feature engineering architecture was proposed, which contained the following components: (i) multiple nested patch division, (ii) deep feature extraction, (iii) multiple feature selection by deploying neighborhood component analysis (NCA), Chi2, and ReliefF selectors, (iv) classification using SVM with 10-fold cross-validation, (v) voted results generation by deploying iterative hard majority voting (IHMV) and (vi) selection of the best vector by a greedy algorithm. Results: Our proposal attained a 91.87% classification accuracy on the collected dataset. This is the best result of our presented framework, which was automatically selected from 70 generated results. Conclusions: The computed classification results and findings demonstrated that monkeypox could be successfully detected using our proposed automated model. [ABSTRACT FROM AUTHOR]

Microbiologically Induced Calcium Carbonate Precipitation (MICP) is a technology for improving soil characteristics, especially strength, that has been gaining increasing interest in literature during the last few years. Although a lot of influencing factors on the result of MICP are known, particle size and shape of the particles remain poorly understood. While destructive measuring of compressive strength or calcium carbonate content are important for the characterization of samples these methods give no insight into the internal structures and pore networks of the samples. X-ray microcomputed tomography (micro-CT) is a technique that is used to characterize the internals of rocks and to a certain degree MICP-treated soils. However, the impact of filtering and image processing of micro-CT Data depending on the type of MICP sample is poorly described in the literature. In this study, single fractions of local quarry were treated with MICP through the ureolytic microorganism Sporosarcina pasteurii to investigate the influence of particle size distribution on calcium carbonate content, unconfined compressive strength and the reduction of water permeability. Additionally, micro-CT was conducted to obtain insights into the resulting pore system. The impact of the Gauss filter und Non-local means filter on the resulting images and data on the pore network are discussed. The results show that particle size has a significant impact on the result of all tested parameters of biosandstone with lower particle size leading to higher strength and generally higher calcium carbonate content. Micro-CT data showed that the technology is feasible to gain valuable insights into the internal structures of biosandstone but the resolution and signal-to-noise ratio remain challenging, especially for samples with particle sizes smaller than 125 µm. Article highlights: Microbiologically Induced Calcium Carbonate Precipitation is more effective for smaller sand particle sizes Smaller particle size of sand lead to higher permeability reduction, higher compressive strength and higher calcium carbonate contents in biosandstone X-ray microcomputed tomography and image analysis is sufficient to characterize the pore system of biosandstone and detect uneven biocemented areas Usage of different digital filters influence resulting images and data from X-ray microcomputed tomography [ABSTRACT FROM AUTHOR]

Fabric pilling can significantly affect the usability and appearance of fabrics; making the evaluation of pilling grades a crucial aspect of textile quality inspection. Currently, manual visual inspection is the mainstream strategy in the industry for assessing pilling, but this method suffers from low efficiency and inconsistent accuracy. Therefore, there is a need for a more precise and efficient method to evaluate textile pilling grades. In this paper, we propose an objective system for assessing fabric pilling grades using a deep learning model in conjunction with a gradient boosting decision tree classification algorithm. The system comprises a multi-light source fabric image acquisition setup, an Attention-enhanced UNet model, and a gradient boosting decision tree classifier. The process begins with capturing fabric images, which are then segmented using the Attention-enhanced UNet model. This model effectively extracts pilling features, including the number of pills, the pilling area, and the pilling coverage ratio. Our experimental results show that this segmentation method outperforms traditional filtering and edge detection algorithms in terms of efficiency. The attention mechanism integrated into the UNet model significantly enhances the segmentation performance, particularly for fabric pilling. By combining the extracted features as input for the fabric evaluation algorithm, our proposed method achieves an impressive classification accuracy of 98.47% in objectively evaluating fabric pilling grades. This approach offers a substantial improvement in both the accuracy and efficiency of fabric pilling grade assessment. [ABSTRACT FROM AUTHOR]

Flame temperature field measurement has always been a key topic in combustion research, which is of great significance for combustion state diagnosis and fuel combustion optimization. Deep learning technology exploits its superior nonlinear ability to rapidly reconstruct the three-dimensional (3D) temperature field of flames from flame light-field images. However, existing algorithms have problems with complex networks, poor noise resistance and low reconstruction accuracy. This paper proposes a lightweight 3D flame temperature field reconstruction algorithm based on an improved MobileNet that combines residuals and spatial attention mechanisms. This method basically achieves a balance between low complexity and high accuracy and has good noise resistance performance. Simulation results show that the average relative error of temperature field reconstruction on the noise-free unimodal flame dataset is only 0.022%, and its computational complexity is only one-tenth of the existing CNN. The noise simulation experiment shows that this method has good noise resistance performance. The maximum relative error and the average relative error at Gaussian white noise standard deviation σ = 0.15 are only 2.91% and 0.27%, respectively. [ABSTRACT FROM AUTHOR]

This study introduces a novel approach that addresses the limitations of existing methods by integrating 2D image processing with 3D point cloud analysis, enhanced by interpretable neural networks. Unlike traditional methods that rely on either 2D or 3D data alone, our approach leverages the complementary strengths of both data types to improve detection accuracy in environments adversely affected by welding spatter and smoke. Our system employs an improved Faster R-CNN model with a ResNet50 backbone for 2D image analysis, coupled with an innovative orthogonal plane intersection line extraction algorithm for 3D point cloud processing. By incorporating explainable components such as visualizable feature maps and a transparent region proposal network, we address the "black box" issue common in deep learning models.This architecture enables a more transparent decision-making process, providing technicians with necessary insights to understand and trust the system's outputs. The Faster-RCNN structure is designed to break down the object detection process into distinct, understandable steps, from initial feature extraction to final bounding box refinement. This fusion of 2D-3D data analysis and interpretability not only improves detection performance but also sets a new standard for transparency and reliability in automated welding systems, facilitating wider adoption in industrial applications. [ABSTRACT FROM AUTHOR]

Deep learning applications have witnessed significant advancements across diverse domains, revolutionizing tasks such as image recognition, disease classification, cancer detection, natural language processing, and autonomous decision-making. However, striking a balance between performance and cost remains a fundamental challenge. Performance, which encompasses accuracy, efficiency, and effectiveness, is crucial for the real-world applicability of deep learning models. Attaining high performance often involves using resource-intensive models with intricate architectures, extensive datasets, and hyperparameter tuning. While these factors lead to improved performance, they also result in increased costs, including computational resources, time, and energy consumption. This research addresses this challenge by proposing a novel low-cost deep learning model based on Sharpened Cosine Similarity (LC-SCS) for breast cancer classification using thermal images. In our study on a DMR-IR dataset, we also utilized pre-trained models: ResNet-101, VGG-16, Inception-V3, ResNet-50, VGG-19, and Xception. The proposed LC-SCS model achieved an impressive accuracy of 94%, trailing just 4% behind the leading VGG-19 model, while maintaining a low computational cost. It achieved recall and precision scores of 0.95 each, with an F1-score of 0.94. The LC-SCS model excels in computational efficiency, demanding only 1.85 GFLOPs and 1.2 million parameters. The memory requirement is minimal at 865.35 MB, and it exhibits a latency of just 0.003 s. Additionally, the CPU execution time for prediction is 3.59 s. Comparatively, the best-performing pre-trained VGG-19 model achieved 98% accuracy but incurred significantly higher costs. The proposed LC-SCS model showcases a remarkable 298.63 times reduction in GFLOPs, a 116.31 times reduction in total parameters, a 4.57 times reduction in memory requirements for model weights, a 5 times reduction in latency, and a 3.33 times reduction in CPU execution time for prediction, thus making it as an exceptionally resource-efficient model, particularly in scenarios with limited computational resources. [ABSTRACT FROM AUTHOR]

Automatic skin lesion subtyping is a crucial step for diagnosing and treating skin cancer and acts as a first level diagnostic aid for medical experts. Although, in general, deep learning is very effective in image processing tasks, there are notable areas of the processing pipeline in the dermoscopic image regime that can benefit from refinement. Our work identifies two such areas for improvement. First, most benchmark dermoscopic datasets for skin cancers and lesions are highly imbalanced due to the relative rarity and commonality in the occurrence of specific lesion types. Deep learning methods tend to exhibit biased performance in favor of the majority classes with such datasets, leading to poor generalization. Second, dermoscopic images can be associated with irrelevant information in the form of skin color, hair, veins, etc.; hence, limiting the information available to a neural network by retaining only relevant portions of an input image has been successful in prompting the network towards learning task-relevant features and thereby improving its performance. Hence, this research work augments the skin lesion characterization pipeline in the following ways. First, it balances the dataset to overcome sample size biases. Two balancing methods, synthetic minority oversampling TEchnique (SMOTE) and Reweighting, are applied, compared, and analyzed. Second, a lesion segmentation stage is introduced before classification, in addition to a preprocessing stage, to retain only the region of interest. A baseline segmentation approach based on Bi-Directional ConvLSTM U-Net is improved using conditional adversarial training for enhanced segmentation performance. Finally, the classification stage is implemented using EfficientNets, where the B2 variant is used to benchmark and choose between the balancing and segmentation techniques, and the architecture is then scaled through to B7 to analyze the performance boost in lesion classification. From these experiments, we find that the pipeline that balances using SMOTE and uses the adversarially trained segmentation network achieves the best baseline performance of 91% classification accuracy with EfficientNet B2. Based on the scaling experiments, we find that optimal performance is reached with the B6 architecture that classifies with a 97% accuracy. Furthermore, the proposed pipeline for lesion characterization outperforms the state of the art performance on the ISIC dataset. [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]

Deep neural networks (DNNs) are popular in image processing but are vulnerable to adversarial attacks, which makes their deployment in security-sensitive systems risky. Adversarial attacks reduce the performance of DNNs by generating adversarial examples (AEs). In this paper, we propose a novel method called Trans-IFFT-FGSM (Transformer Inverse Finite Fourier Transform Fast Gradient Sign Method) to generate adversarial examples. Unlike others, we apply multiple steps, adding imperceptible perturbation and saving input noise information to create strong AEs, while emphasizing simplicity, efficiency, robustness through iterations, and analytical precision on specific models. We evaluate and compare perturbation generated by Trans-IFFT-FGSM and other attack methods, including FGSM, PGD, DeepFool, and C &W on ImageNet and MNIST, and evaluation results suggest that Trans-IFFT-FGSM achieves a high attack success rate (ASR) and attack accuracy. In addition, we compare Trans-IFFT-FGSM and other attack methods under the existence of a defense method, which denoises the AEs generated by these methods, and the evaluation results also suggest Trans-IFFT-FGSM outperforms other methods. [ABSTRACT FROM AUTHOR]

The field of computer vision has witnessed significant interest in the area of salient object recognition. The utilization of this technology becomes advantageous in various applications, such as image segmentation, image attention retargeting, image cropping, and image understanding. The primary challenges encountered by underwater images pertain to diminished contrast, distorted colors, and an overall suboptimal visual appearance. The present study introduces an innovative and resilient method for detecting saliency in underwater photos. The RGB picture input undergoes enhancement and is subsequently subjected to gradient domain filtration. The filtered image is subjected to a pyramid decomposition. The computation of saliency is performed on three distinct scales using a process consisting of two parallel phases. In the first stage, saliency is calculated by employing cellular automata at various scales. Initially, the filtered image is subjected to a decomposition process into superpixels. Subsequently, the k-means clustering technique is employed to calculate matrices that represent color dissimilarity and geodesic distance. The application of cellular automata follows the fusion of the matrices. The map undergoes additional optimization and filtering processes in order to achieve saliency at a specific scale. The computation of a main saliency map involves the process of scale integration. In the second stage, the computation of saliency is achieved by applying a homomorphic filter to each scale. The natural logarithm of the filtered image is computed. The spatial domain of the image undergoes a transformation to the frequency domain. The Butterworth high-pass filter is utilized. The frequency domain of the image is subsequently converted back to its spatial domain. The computation of the inverse logarithm of the image is performed. The guided filter is employed to acquire saliency on a specific scale. The computation of the secondary saliency map involves the process of scale integration. The final output is obtained by applying multiplicative fusion to both primary and secondary saliency maps. Utilizing cutting-edge methodologies, a thorough evaluation that includes both qualitative and quantitative analyses evaluates the effectiveness of the suggested strategy. In comparison to alternative state-of-the-art methodologies, the findings indicate that the suggested methodology exhibits high levels of precision and dependability. [ABSTRACT FROM AUTHOR]

Black ice is responsible for dangerous road-related incidents that can cause collisions and harm vehicle drivers and pedestrians. Visual examination and weather forecasts are two standard traditional methods for detecting black ice on roads, but they are often inaccurate and may not deliver the vehicle driver with up-to-date information on road conditions. The evolution of Industry 4.0 enabling technologies such as wireless sensor network (WSN), Internet of Things (IoT), cloud computing, and machine learning (ML) has been capable of detecting events in real time. This study aims to analyse the integration of the WSN, IoT, ML, and image processing for black ice detection. The qualitative research method is followed in this study, where the problems of black ice detection are studied. Following this, the role of Industry 4.0 enabling technologies is analyzed in detail for black ice detection. According to the study, we can detect black ice using different methods, but some methods need to be refined if we talk about the prediction. By merging different technologies, we can improve the overall architecture and create an algorithm that works with images and physical variables like temperature, humidity, due point, and road temperature, which were responsible for black ice formation, and predict the chances of black ice formation by training the system. [ABSTRACT FROM AUTHOR]

This study aims to evaluate the applicability of the high-resolution WaveFront Phase Imaging Sensor (WFPI) in eyes with Fuchs' Endothelial Corneal Dystrophy (FECD) through qualitative and quantitative analysis using a custom-designed Automatic Guttae Detection Method (AGDM). The ocular phase was measured using the t · eyede aberrometer and then was processed to obtain its High-Pass Filter Map (HPFM). The subjects were pathological and healthy patients from the Fundación Jiménez-Díaz Hospital (Madrid, Spain). The AGDM was developed and applied in pupils with 3 and 5 mm of diameter. A set of metrics were extracted and evaluated like the Root-Mean-Square error (RMS), Number of guttae, Guttae Area, and Area of Delaunay Triangulation (DT). Finally, a Support Vector Machine (SVM) model was trained to classify between pathological and healthy eyes. Quantitatively, the HPFM reveals a dark spots pattern according to the ophthalmologist's description of the slit-lamp examination of guttae distribution. There were significant statistical differences in all the metrics when FECD and Healthy groups were compared using the same pupil size; but comparing both pupil sizes for the same group there were significant differences in most of the variables. This sensor is a value tool to objectively diagnose and monitor this pathology through wavefront phase changes. [ABSTRACT FROM AUTHOR]

Every year, more than two million Muslim pilgrims from all over the world visit Mecca to perform Hajj worship. The challenge of location identification in dense crowds is significant, leading to dangerous consequences such as injury or loss. Existing works for person localization remain challenged, especially in crowded places like Mecca during Hajj. In this work, we propose a novel location identification method using image pre-processing and different machine learning classifiers, with the creation of a new image dataset for hotspot locations in Mecca. Image pre-processing algorithms are applied to enhance the geographic information present in the images, and the obtained features are classified using CNN, ANN, and SVM classifiers. Extensive evaluations reveal that the proposed pre-processing algorithm with CNN achieves the best localization performance with an accuracy of 90%, followed by ANN with an accuracy of 84%, and SVM with an accuracy of 80.50%. Without pre-processing, the accuracies are significantly lower: 63% for CNN, 73% for ANN, and 71.50% for SVM. In addition, the proposed approach was compared with other deep learning models, VGG16, AlexNet and ResNet50 on our dataset achieving an accuracy of 61%, 65% and 62%, respectively. Results demonstrate the effectiveness of our proposed method in comparison with other deep transfer learning methods on a small dataset, offering promising solutions for crowded place navigation. [ABSTRACT FROM AUTHOR]

With the increasing complexity of digital systems, managing power dissipation and energy consumption in digital circuits, particularly in emerging embedded systems for artificial intelligence and signal processing applications, has become a challenging issue. The emerging paradigm of approximate computing offers the potential to reduce energy consumption and enhance speed by trading accuracy. This paper introduces several 4-bit approximate multipliers and outlines a systematic approach for constructing extended bit-length multipliers by leveraging lower bit-length counterparts. Our evaluations confirm that, in comparison with state-of-the-art approximate multipliers, the proposed multipliers demonstrate energy reduction of up to 50, 76, and 83% in 8-bit, 16-bit, and 32-bit multipliers, respectively. Additionally, the reduction in energy-delay product (EDP) reaches up to 86, 93, and 97%, correspondingly. The efficiency of the proposed approximate multipliers has been explored and confirmed in executing various image processing algorithms, a regression model developed for stock price prediction, and in executing quadrature amplitude demodulation. [ABSTRACT FROM AUTHOR]

This paper investigates the initiation and evolution of polygonal crack patterns in desiccating soil layers with varying thicknesses and grain sizes, both with and without a secondary saturated sand layer placed beneath the soil layer. Single-layer experiments involved soil samples within a specific range of grain sizes (0–100, 100–300, 300–500, and 500–850 μm), maintaining soil layer depths at 10, 20, 30, 40, 50, and 60 mm. In double-layer experiments, a saturated sand layer was introduced below the soil layer to check the role of relative layer thickness on crack patterns. Time-lapse photography captured surface crack development during desiccation, allowing for the measurement of geometric parameters like crack width, crack intensity factor (CIF), and intersection angle (CIA). The single-layer models indicate an increase in crack width and CIF with greater layer thicknesses, while CIA decreases with increased layer thickness. Additionally, experiments with finer grain sizes exhibit relatively wider cracks, along with higher CIF and CIA. In double-layer models with varying thicknesses of individual layers, crack growth is found to be independent of the upper soil layer's thickness. Instead, crack propagation is controlled by the lower sand layer, as the supply of water from the lower sand layer to the upper soil layer facilitates prolonged desiccation, resulting in larger CIF values. Using digital image processing and the box-counting method, we calculated the fractal dimensions (D) of the cracks were calculated. D demonstrates positive relationships with grain size in both single- and double-layer experiments, suggesting a self-similar evolution of crack patterns in the models with coarse-grained soil. Research highlights: The role of basal water-rich sand layer in the growth of mud-cracks is investigated. Layer thickness and grain size are two additional variables during the experiments. Single mud-layer experiments were used to compare the double-layer experiments. Basal wet-sand delays desiccation, resulting in a higher crack intensity factor. Finer grains produce denser cracks irrespective of the imposed variables. Low fractal dimensions in course-grained soils imply a self-similar crack pattern. [ABSTRACT FROM AUTHOR]

In image processing, computer vision algorithms are applied to regions bounded by closed contours. These contours are often irregular, poorly defined, and contain holes or unavailable areas inside. A common problem in computational geometry includes finding the k-sided polygon (k-gon) of maximum area or maximum perimeter inscribed within a contour. This paper presents a generic method to obtain user-defined polygons within a region. Users can specify the number k of sides of the polygon to obtain. Additionally, users can also decide whether the calculated polygon should be the largest in area or perimeter. This algorithm produces a polygon or set of polygons that can be used to segment an image, allowing only relevant areas to be processed. In a real-world application, the validity and versatility of the proposed method are demonstrated. In addition, the source code developed in Java and Python is available in a GitHub repository so that researchers can use it freely. [ABSTRACT FROM AUTHOR]

To communicate with one another hand, gesture is very important. The task of using the hand gesture in technology is influenced by a very common way humans communicate with the natural environment. The recognizing and finding pose estimation of hand comes under the area of hand gesture analysis. To find out the gesturing hand is very difficult than finding the another part of the human body because the hand is smaller in size. The hand has greater complexity and more challenges due to differences between the cultural or individual factors of users and gestures invented from ad hoc. The complication and divergences of finding hand gestures will deeply affect the recognition rate and accuracy. This paper emphasizes on summary of hand gestures technique, recognition methods, merits and demerits, various applications, available data sets, and achieved accuracy rate, classifiers, algorithm, and gesture types. This paper also scrutinizes the performance of traditional and deep learning methods on dynamic hand gesture recognition. [ABSTRACT FROM AUTHOR]

Access to large image volumes through camera traps and crowdsourcing provides novel possibilities for animal monitoring and conservation. It calls for automatic methods for analysis, in particular, when re-identifying individual animals from the images. Most existing re-identification methods rely on either hand-crafted local features or end-to-end learning of fur pattern similarity. The former does not need labeled training data, while the latter, although very data-hungry typically outperforms the former when enough training data is available. We propose a novel re-identification pipeline that combines the strengths of both approaches by utilizing modern learnable local features and feature aggregation. This creates representative pattern feature embeddings that provide high re-identification accuracy while allowing us to apply the method to small datasets by using pre-trained feature descriptors. We report a comprehensive comparison of different modern local features and demonstrate the advantages of the proposed pipeline on two very different species. [ABSTRACT FROM AUTHOR]

Most commercially available eye-tracking devices rely on video cameras and image processing algorithms to track gaze. Despite this, emerging technologies are entering the field, making high-speed, cameraless eye-tracking more accessible. In this study, a series of tests were conducted to compare the data quality of MEMS-based eye-tracking glasses (AdHawk MindLink) with three widely used camera-based eye-tracking devices (EyeLink Portable Duo, Tobii Pro Glasses 2, and SMI Eye Tracking Glasses 2). The data quality measures assessed in these tests included accuracy, precision, data loss, and system latency. The results suggest that, overall, the data quality of the eye-tracking glasses was lower compared to that of a desktop EyeLink Portable Duo eye-tracker. Among the eye-tracking glasses, the accuracy and precision of the MindLink eye-tracking glasses were either higher or on par with those of Tobii Pro Glasses 2 and SMI Eye Tracking Glasses 2. The system latency of MindLink was approximately 9 ms, significantly lower than that of camera-based eye-tracking devices found in VR goggles. These results suggest that the MindLink eye-tracking glasses show promise for research applications where high sampling rates and low latency are preferred. [ABSTRACT FROM AUTHOR]

As the population in cities increases, traffic problems have emerged, especially at intersections with high traffic density. Increasing traffic density leads to longer transportation times, higher fuel consumption, and elevated levels of environmental pollution. Various techniques have been employed to decrease traffic congestion. In order to apply these methods, the degree of traffic density must first be determined. This is typically done through vehicle counting studies in the field using camera images. However, manually counting vehicles from camera images is a very detailed process. Therefore, various automated methods based on image processing techniques are preferred today to perform these operations faster and more accurately. In this study, we designed virtual zones using different vehicle counting methods at intersections based on image processing techniques. We obtained vehicle count data from four methods, including manual counting and three methods based on image processing techniques. We evaluated the accuracy of the counting results using transportation engineering parameters such as density and traffic volume. Additionally, we modeled the signalized intersection in the AIMSUN simulation program. The study found that the "New Type Virtual Zone" method resulted in vehicle counts that were 95% accurate, and the average success rate of the AIMSUN simulation analysis results performed with this data was 83.71% accurate. [ABSTRACT FROM AUTHOR]

The multilevel threshold technique of image segmentation is a popular and intriguing domain in the field of image vision and has received a lot of attention in several image processing applications due to its use in numerous image applications for assisting with a variety of problems. The key issue in this domain is figuring out the optimal number of thresholds and their values due to the limitations of conventional algorithms, such as fixed threshold values, a lack of adaptability, manual parameter setting, and a lack of contextual information. In order to deal with this problem, a new multilevel thresholding (MLT) algorithm (IP-MH-MLT) has been proposed in this paper. It is based on the image partitioning (IP) approach and has a few parameters that are computed using any metaheuristic (MH) technique, and the remaining parameters are evaluated through image characteristics. In this paper, for the metaheuristic parameter, a swarm-based metaheuristic called Grey Wolf Optimizer (GWO) is taken into consideration due to its numerous features, such as simplicity and ease of implementation, efficient convergence speed, and limited computational complexity. The performance of the proposed algorithm has been validated on a set of fifteen benchmark images using various thresholds and compared quantitatively in terms of a number of performance metrics, including structural similarity index measure (SSIM) and peak signal-to-noise ratio (PSNR), with eight different metaheuristic algorithms. In order to examine the proposed algorithm qualitatively, Friedman ranking tests are carried out on the proposed algorithm over other comparable algorithms. The results demonstrate the effectiveness and competitive performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Modern μ -CT scans offer non-destructive three-dimensional microstructure analysis of various materials. Despite small density contrasts, this technology also accurately captures the complex network structure of paper and paperboard. It provides detailed insight into the fiber orientation distribution in paper, which was previously unattainable with existing measurement methods. The image analysis results are applicable for numerical fiber network models, and simulations based on these models can significantly improve our understanding of the microstructural influence on macrostructural paper properties, such as strength, stiffness, and curl tendency. This work presents a new method for investigating the microstructural fiber orientation in paperboard. Orientation tensors obtained from μ -CT scan image processing are extracted and evaluated. Based on the direction of the orientation tensor's first eigenvector, the fiber orientation distribution is determined and then approximated by periodic and non-periodic probability density functions (PDFs). In this study, 40 commercial paperboard samples, each consisting of three plies, were analyzed. From a given set of commonly used PDFs, the best fitting ones have been identified based on their original location in the paper roll and on their ply. It was found that the von Mises PDF provided the most accurate representation of fiber orientation distribution in the middle ply, while the Elliptical PDF was the most suitable for the outer plies. Moreover, a more pronounced fiber orientation anisotropy was observed at the edges of the paper roll than in its center. The best PDFs and their function parameters are provided to allow for direct usage in numerical microstructure models of paperboard, thus enhancing their representativeness. [ABSTRACT FROM AUTHOR]

Potholes are a significant hazard, causing severe damage to vehicles and potentially leading to fatal accidents. The automatic detection of potholes is crucial for timely maintenance and reducing these risks. This research evaluates the performance of three pre-trained Convolutional Neural Network models—ResNet 50, ResNet 18, and MobileNet—in classifying pavement images. Initially, the models distinguish between images containing potholes and those without (Potholes vs. Normal). Subsequently, they categorize pavement images into three classes: Small Pothole, Large Pothole, and Normal. Pavement images were captured from two different heights: 3.5 feet (waist height) and 2 feet. Among the models, MobileNet v2 demonstrated the highest accuracy of 98% for detecting potholes. For images taken at 2 feet, the classification accuracies for large potholes, small potholes, and normal pavement were 87.33%, 88.67%, and 92%, respectively. For images taken from waist height, the accuracies increased to 98.67%, 98.67%, and 100%, respectively. The study highlights the potential of these models in enhancing road safety through reliable and automated pothole detection, providing valuable insights for infrastructure maintenance. Article Highlights: Potholes are dangerous and can cause serious vehicle damage. Train CNN models for the classification of potholes These CNNs, for different sizes of potholes can be used for the rehabilitation of potholes. Among all CNN modes, MobileNet v2 achieves 98% accuracy in detecting the presence of a pothole. [ABSTRACT FROM AUTHOR]

Background: To protect greenhouse crops from pests, millions of liters of toxic solutions are used each year. To ensure food security, it is very important to reduce the consumption of these toxins. Monitoring the pesticide residue in plants and fruits is a vital research topic. In this study, the magnitude and percentage of the contaminated area with the pesticide were evaluated in tomato fruit after treating tomatoes using a High-Voltage Electric Field (HVEF). The parameters related to the HVEF for monitoring pesticide residue changes have been assessed to simulate variable scenarios of the application of pesticides on land. Using an image processing technique, a percentage of pesticide residue reduction was measured in MATLAB. Results: The effect of independent parameters was significantly different, no interaction was significant and it revealed that the penetration of the pesticide does not affect the performance of the HVEF to reduce the residue in tomato samples. Conclusion: The application of HVEF for 10 min is the best performance of the device to reduce pesticide residues. The percentage of pesticide residue reduction, which is an efficient way to monitor pesticide residues in tomatoes utilizing image processing techniques, can also be established. [ABSTRACT FROM AUTHOR]

This study aims to obtain a sustainable substitute for man-made fibers using Sambucus ebulus L. as a reinforcement in the polymer composites. With this purpose, an extensive characterization of Sambucus ebulus L. stem fiber was performed. The chemical composition was obtained approximately as 76 wt.% cellulose, 66 wt.% hemicellulose, and 4.47 wt.% lignin. Scanning electron microscope analysis proved that the fibers are nonuniform and have rough surface along the length. The elemental composition was found as 58.67% carbon and 40.34% oxygen with energy-dispersive X-ray analysis. Furthermore, the fiber diameter was estimated as 759.41 µm with image processing from scanning electron microscope graphs of fibers. X-ray diffraction analysis verified the existence of crystalline cellulose having 72% crystalline index with 3.9 nm crystalline size. Functional groups of the fiber were determined with Fourier transform infrared analysis. A single fiber tensile test found out that the fibers have the mean tensile strength, elongation at break, and Young's modulus of 90.6 MPa, 6.45%, and 1.46 GPa, respectively. Moreover, the obtained fibers are thermally durable to 456.98 °C with thermogravimetric analysis. The results indicate that Sambucus ebulus L. stem fibers have the comparable properties to other natural cellulosic fibers. It is suggested to use Sambucus ebulus L. stem fiber in composites as reinforcement to substitute for man-made fibers. [ABSTRACT FROM AUTHOR]

Hot forging is one of the common manufacturing processes for producing brass workpieces. However forging produces flash which is a thin metal part around the desired part formed with an excessive material. Using robots with vision system to manipulate this workpiece has encountered several challenging issues, e.g. the uncertain shape of flash, color, reflection of brass surface, different lighting condition, and the uncertainty surrounding the position and orientation of the workpiece. In this research, Mask region-based convolutional neural network together with image processing is used to resolve these issues. The depth camera can provide images for visual detection. Machine learning Mask region-based convolutional neural network model was trained with color images and the position of the object is determined by the depth image. A dual arm 7 degree of freedom collaborative robot with proposed grasping strategy is used to grasp the workpiece that can be in inappropriate position and pose. Eventually, experiments were conducted to assess the visual detection process and the grasp planning of the robot. [ABSTRACT FROM AUTHOR]

In today's urban development, Earth Pressure Balance (EPB) Tunnel Boring Machines (TBMs) play a vital role. It's crucial to design a comprehensive monitoring system to control surface settlement and prevent damage to surface structures. This study focuses on creating prediction models for estimating ground surface settlement. Two soft computing techniques, namely ANN-CFB and ANN-BP, were used for this purpose. The models were validated using operational data from the Qom metro Line A, specifically the section between A14 and A10 stations. Additional input parameters were incorporated using an image processing approach to include soil properties for each segment. As a result, the most accurate ANN technique was employed to predict ground surface settlements for the mentioned project. The correlation coefficients for training, testing, validation, and the overall result were found to be 0.99439, 0.97873, 0.96381, and 0.98824, respectively. Through sensitivity analysis, the study explored the connections between different parameters and ground surface settlement. The outcomes reveal strong agreement between predicted values and real data. Notably, the parameter 'cutter head torque' exhibited the highest impact on surface settlement (8.48%), while 'Pressiometric Modulus (Ep)' had the least impact (4.24%). [ABSTRACT FROM AUTHOR]

Coastal marsh ecosystems are changing and being lost at a rapid rate around the world. One of the fastest rates of land loss, specifically coastal marsh habitat, occurs in Louisiana on the Northern Gulf Coast of the USA. To address this issue, state and federal agencies have undertaken massive wetland restoration efforts to preserve and restore coastal marsh habitats in Louisiana. For these efforts to be successful in the long-term, it is critical to understand what methodologies and techniques result in resilient restoration projects. However, traditional methods to monitor restoration success rely on labor intensive field measurements that are often limited in scope, difficult to maintain, and underfunded. Recent technological developments with uncrewed aircraft systems (UASs) and image processing have substantially improved the ability of restoration practitioners to use off-the-shelf UASs and cameras to map projects. We present a streamlined method using a commercially available drone with a high-resolution red, green, blue (RGB) camera to assess the effects of wetland restoration and integrate more modern tools into evaluation approaches. We conducted drone flights at restored brackish marshes of various ages using a space for time substitution with the goal of understanding the long-term success of marsh restoration. We observed that created marshes had higher land to water ratios than natural marshes. This finding suggests that these restored areas were gaining and maintaining elevation after approximately 10 years. Our method shows that drone surveys offer low-cost, minimally invasive methods for evaluating restored wetlands and ultimately tell us more about ecosystem function through realistic site-level habitat configurations. [ABSTRACT FROM AUTHOR]

Shoreline detection and estimation of changes is a well-established concept in the field of coastal zone management. Recent technological advancements in the form of machine learning (ML) have transformed shoreline detection methodologies. However, due to the constantly changing nature of coasts, identification of the boundary between land and ocean has become an intricate process. In this particular investigation, long-term changes in shoreline along the lower segment of the Indian Sundarbans Delta (ISD) are estimated by employing Landsat sensor's optical imageries spanning 28 years from 1995 to 2023. A fully automated approach involving support vector machine (SVM) classification algorithm with a 99.5% accuracy and zero-crossing edge detection algorithm for shoreline extraction from optical imagery has been proposed. The implementation stage of shoreline extraction utilizes Google Earth Engine's cloud platform. In contrast, subsequent analysis to calculate shoreline changes conforms to the Digital Shoreline Analysis System (DSAS v.5), an extension of ArcGIS Desktop's functionality. This research article examines the long-term shoreline changes in the Indian Sundarban Delta. The study uses three statistical measures: end point rate (EPR), linear regression rate (LRR), and net shoreline movement (NSM). EPR analysis shows significant erosion on Kanak Island and Bhangaduni Island up to 88.21 m. LRR statistics reveal negative trends on the eastern side. NSM analysis highlights maximum accretion in the northwestern part of the delta. This study offers valuable insights into dynamic coastal processes in the area. [ABSTRACT FROM AUTHOR]

Learning-based multi-view stereo (MVS) has shown great promise in the field of 3D reconstruction. However, existing MVS methods suffer from fixed receptive field sizes during feature learning. This issue leads to information loss and affects the understanding of the geometric structure of the scene, posing a challenge to the reconstruction quality of regions with complex geometric structures and lighting conditions. Therefore, we propose DAR-MVSNet, which consists of a dual-attention-guided feature pyramid network (DA-FPN) and a 3D residual U-net module (3D-RUM). DA-FPN includes two modules: attention-based context extraction module (ACEM) and self-attention-based module (SAM). ACEM is proposed to dilate the receptive field and preliminarily filter deep features through multiple dilated convolutions, spatial and channel attention. To further eliminate redundant characteristics, SAM is proposed to enhance the representation capability of depth features. Moreover, 3D-RUM is designed to enhance feature transfer and information flow, thereby addressing the problem of severe global feature information loss. This article demonstrates the effectiveness of DAR-MVSNet through an extensive series of experiments. The result on the DTU dataset and Tanks and Temples benchmark in comparison to state-of-the-art MVS methods verify its superior performance. [ABSTRACT FROM AUTHOR]

Edge-aware image smoothing refers to the removal of details with edges preserved. It is an essential topic in the field of image processing and computer graphics. In this paper, in order to achieve better edge preservation than the existing models, we propose a robust edge-preserving image filtering method based on a complementary weighting scheme. Both isotropic and anisotropic weights are involved in our model to adapt the fidelity and the regularization terms. To efficiently solve the proposed model, we introduce an effective algorithm based on additive half quadratic minimization, alternating direction of multipliers, and Fourier domain optimization strategies. We experimentally validate the proposed filter on several low-level vision tasks. Both quantitative and qualitative experimental results show significant superiority of our proposed filter compared to existing techniques. Furthermore, the filter exhibits high efficiency and is able to process 720P color images (over 10 fps) in real-time on an NVIDIA RTX 3070. Therefore, it is practical for real-world applications. [ABSTRACT FROM AUTHOR]

Cracks are a common problem in concrete surfaces. With the continuous optimization of machine vision-based inspection systems, effective crack detection and recognition is the core of the entire system. In this study, support vector machine (SVM) was used to distinguish cracks from other regions. To complete the recognition system of the SVM, a framework consisting of an image processing and recognition model was proposed. An algorithm combining the Prewitt operator with the Otsu threshold was proposed for image segmentation. The binary image processed by the new algorithm combined with mathematical morphology can result in a more complete crack zone and fewer interference regions. After the initial parameter extraction, most of the impurity areas were screened by preliminary discrimination, removing 99% of the impurities. This processing step ensured the balance and effectiveness of the samples. To establish an automatic identification model based on SVM with a radial basis function, compactness, occupancy rate, and length–width ratio were selected as input parameters after comparing these three features with all six features of the crack. The recognition accuracy of this system reaches 97.14%, demonstrating that the proposed method is effective and satisfies practical requirements. [ABSTRACT FROM AUTHOR]

Convolutional neural networks (CNNs) are used to solve pattern classification problems. As this algorithm is based on establishing a relationship between an image-shaped input and its related output through the CNN structure, the training stage is a significant process in their working basis. This study develops a new-fangled and explainable algorithm to train CNNs. The input filters in the convolution layers are parameterized to keep the filter structure, implementing traditional and explainable image processing filters within the network topology. A back-propagation scheme updates the parameters in the filters and the fully connected section of the CNN. Several traditional image filters (Sobel, averaging, Gaussian, and directional, among others) are used in CNN with a learning strategy that keeps their kernel structures. The method implies that the training of these networks is applied to a single parameter instead of all coefficients in the filters, reducing the uncertainty about how each filter performs the image analysis in CNN. This approach was compared with traditional CNNs considering the analysis of the computational cost (measured in terms of time and floops required for training) and their accuracy results. Three image databases were used to evaluate the proposed algorithm. Using a cross-validation methodology, the new training algorithm based on the filter parameterization achieved higher accuracy (93.7% vs. 91.3% in average). The new algorithm got similar results regarding the computational cost compared to traditional methods. This characteristic makes the proposed training methodology an appropriate option to classify images with more explainable processing at the convolution layers. [ABSTRACT FROM AUTHOR]

In this paper, a framework of obstacle avoidance algorithm applied to power line damage safety distance detection is constructed, and its overall architecture and key processes are described in detail. The system design covers three core modules: visual data acquisition and preliminary processing, accurate target recognition and distance measurement, and system error analysis and correction. In the visual data processing chain, we deeply analyze every step from image acquisition to preprocessing to feature extraction, aiming to enhance the adaptability of applications to complex scenes. The target recognition and distance estimation part integrates advanced technology of deep learning to improve the reliability of recognition accuracy and distance estimation. In addition, many common error sources, such as system bias, parallax discontinuity, fluctuation of illumination conditions, etc., are discussed in depth, and corresponding correction strategies are proposed to ensure the accuracy and stability of the system, which provides powerful technical support for achieving efficient and accurate safety monitoring. Specifically, by carefully adjusting the learning rate, convolution kernel size, batch size, pooling layer type, and number of hidden layer nodes, we succeeded in improving the overall accuracy from the initial average of 92.4–95%, and the error rate decreased accordingly. [ABSTRACT FROM AUTHOR]

Fabric Defect (FD) detection has played an important role in the process of fabric production. FDs lead to a reduction in cloth prices, causing a significant loss of 45% to 65% for the cloth manufacturer. The detection of FDs becomes a challenging task in the fabric industry due to its complicated shapes and a considerable amount of FDs. Thus, a new detection technique, which has great detection speed and accuracy, is desired to replace manual work. With the emergence of Convolution Neural Networks (CNN) and the development of machine vision and Deep Learning (DL), various detection techniques, combining the benefits of DL and machine vision, have emerged, which replace manual and image processing approaches. The study presents an Automated Fabric Defect Detection using a Hybrid Particle Cat Swarm Optimizer with a Deep Learning (AFDD-HPCSODL) algorithm. AFDD-HPCSODL method aims to detect and categorize the existence of defects in fabric production. The objective is to automate and improve the performance of fabric defect detection, addressing the complexity of different defect patterns and characteristics in textile images. In the presented AFDD-HPCSODL technique, two stages of data pre-processing take place namely contrast enhancement and bilateral filtering (BF)-based noise elimination. Besides, a fusion of feature extraction processes is carried out using Inception V3 and EfficientNetB3. Moreover, the Attention Convolutional Long Short-Term Memory (ACLSTM) network is exploited for the detection and classification of FDs. Furthermore, the HPCSO technique carries out the hyperparameter selection of the ACLSTM model. Finally, the Root Mean Square Propagation (RMSProp) optimizer with the NestNet model is utilized for the defect segmentation process. The stimulation validation of the AFDD-HPCSODL method is studied on the two FD datasets. Widespread comparative results emphasized the improved performance of the AFDD-HPCSODL technique on FD classification with maximum accuracy of 97.92% and 98.89% on the Kaggle FD dataset and the ZJU-Leaper dataset, correspondingly. [ABSTRACT FROM AUTHOR]

Lung cancer is currently the leading cause of cancer-related death. To lower the mortality from lung cancer, early detection is crucial. A precise and effective method of diagnosis by medical professionals is necessary for early detection of lung-related cancers in order to maximize the success rate of treatment. It is particularly difficult to catch it in the early stages of dissemination because there are no symptoms in the early stages. Using specific image processing ideas on computed tomography (CT), we can identify the tumor state in the early phases of spread and diagnose this at an early level. The early detection is crucial for curing and restraining the spread of uncontrolled malignant cells. The best outcomes for accurately identifying malignant tumors in their early stages come from combining the effects of morphological filtering and a contour-based picture segmentation technique. The segmented lung CT scan pictures that were malignant were separated from the rest of the testing dataset by a support vector machine (SVM) classifier in the subsequent analysis phase. Various image quality and performance metrics, such as accuracy, sensitivity, precision and specificity are used to validate the proposed technique. [ABSTRACT FROM AUTHOR]

Emerging Industry 5.0 designs promote artificial intelligence services and data-driven applications across multiple places with varying ownership that need special data protection and privacy considerations to prevent the disclosure of private information to outsiders. Due to this, federated learning offers a method for improving machine-learning models without accessing the train data at a single manufacturing facility. We provide a self-adaptive framework for federated machine learning of healthcare intelligent systems in this research. Our method takes into account the participating parties at various levels of healthcare ecosystem abstraction. Each hospital trains its local model internally in a self-adaptive style and transmits it to the centralized server for universal model optimization and communication cycle reduction. To represent a multi-task optimization issue, we split the dataset into as many subsets as devices. Each device selects the most advantageous subset for every local iteration of the model. On a training dataset, our initial study demonstrates the algorithm's ability to converge various hospital and device counts. By merging a federated machine-learning approach with advanced deep machine-learning models, we can simply and accurately predict multidisciplinary cancer diseases in the human body. Furthermore, in the smart healthcare industry 5.0, the results of federated machine learning approaches are used to validate multidisciplinary cancer disease prediction. The proposed adaptive federated machine learning methodology achieved 90.0%, while the conventional federated learning approach achieved 87.30%, both of which were higher than the previous state-of-the-art methodologies for cancer disease prediction in the smart healthcare industry 5.0. [ABSTRACT FROM AUTHOR]

Accurate diagnosis of plant diseases by the assessment of pathogen presence to reduce disease-related production loss is one of the most fundamental issues for farmers and specialists. This will improve product quality, increase productivity, reduce the use of fungicides, and reduce the final cost of agricultural production. Today, new technologies such as image processing, artificial intelligence, and deep learning have provided reliable solutions in various fields of precision agriculture and smart farm management. In this research, microscopic image processing and machine learning have been used to identify the spores of four common tomato fungal diseases. A dataset including 100 microscopic images of spores for each disease was developed, followed by the extraction of the texture, color, and shape features from the images. The classification results using random forest revealed an accuracy higher than 98%. Besides, as a reliable feature selection algorithm, the butterfly optimization algorithm (BOA) was used to detect the effective image features to identify and classify diseases. This algorithm recognized image textural features as the most effective features in the diagnosis and classification of disease spores. Considering only the eight most effective features selected with BOA resulted in an accuracy of 95% in disease detection. To further investigate the performance of the proposed method, its accuracy was compared with the accuracies of convolutional neural networks and EfficientNet as two reliable deep learning algorithms. Not only the prediction accuracy of these methods was not favorable (65 and 83.55%, respectively), they were very time-consuming. According to the findings, the proposed framework has high reliability in disease diagnosis and can help in the management of tomato fungal diseases. [ABSTRACT FROM AUTHOR]

The usage of digital devices has increased across the world due to global digitalization. The global digitalization concept arises due to the easy accessibility of the internet around the world. One of the important data that contributes a large amount to the internet is the image data. Images are used in various fields for advertising the products, expressing the individual's point of view, describing the diseases in the medical field, etc. Hence a large number of images need to be transformed into digital images to exchange them in the digital images. The process of converting a normal image into a digital image and then extracting information from it is known as image processing. The existing techniques for image processing exhibit issues like high error rates and poor image processing performance. To overcome these issues, this paper proposes an Attention Long Short-term Domain Adversarial Crossover Orchard (ALSDA-CO) algorithm. The proposed method is highly effective in image processing approach and extracted varied class labels clearly that are not visible effectively. The weight functions of each image are trained effectively and optimize the hyperparameters with best solution. To evaluate the performance of this algorithm, the Fruits 100 dataset is used. The proposed algorithm showed better performance than the existing methods such as pre-trained CNN, LS-SVM, MobileNetV2-LSTM, and DCNN in terms of all the performance measures used. The proposed algorithm attained 98.5% accuracy, 98.9% recall, 98.7% precision, and 98.5% F1 score in image processing. Compared to existing approaches the attained range of accuracy outperformed 1.1% of proposed method than existing approaches. The proposed algorithm also exhibited a reduced error rate and enhanced image processing. [ABSTRACT FROM AUTHOR]

Efficient forecasting of solar flares is of significant importance for better risk prevention. Currently, there is relatively rare research on multi/four-classification of flares, and the influence of the number of time steps and data feature dimensions on the prediction performance of multi-class models has not been considered. In this study, we utilize the Space-weather HMI Active Region Patch (SHARP) data to develop two categories of models for multiclass flare prediction within 24 hr, including direct output four-classification models and four-classification models using a cascading scheme. The former encompasses Random Forest (RF) model, Long Short-Term Memory (LSTM) model, and Bidirectional LSTM (BLSTM) model, while the latter includes BLSTM Cascade (BLSTM-C) model and BLSTM Cascade with Attention Mechanism (BLSTM-C-A) model. These two categories of models are employed to contrast the impact of different numbers of time steps and the predictive performance in solar flare multi/four-classification. Additionally, we conduct, for the first time, feature importance analysis for multi/four-classification solar flare prediction using deep learning models. The main results are as follows: (1) As the number of time steps increases, the True Skill Statistic (TSS) scores of the four deep learning models improve, showing an overall upward trend in predictive performance. The models achieve their optimal performance when the number of time steps reaches 120. (2) Among the direct output four-class models, deep learning models (LSTM and BLSTM) outperform traditional machine learning model (RF). In both multi-class and binary-class predictions using deep learning, the BLSTM-C model performs better than other deep learning models (LSTM, BLSTM, and BLSTM-C-A). (3) In the feature importance analysis, the top-ranked important features include SAVNCPP and R_VALUE, while the least important features include SHRGT45 and MEANPOT. [ABSTRACT FROM AUTHOR]

Quantum edge detection offers a promising avenue for real-time image analysis, addressing constraints faced by classical algorithms. However, existing quantum edge detection methods often rely on classical edge detection operators, leading to the loss of intricate edge details, especially in high-resolution images. Here, we present a novel quantum image edge detection algorithm. Our approach involves transforming the image into the wavelet domain through wavelet transform, performing edge detection, and obtaining the edge image via inverse wavelet transform. This innovative method not only mitigates edge information loss but also enhances precision in delineation. Through comprehensive simulations on a classical computing platform, employing peak signal-to-noise ratio (PSNR) and Edge Preservation Index (EPI) evaluations, our proposed scheme demonstrates superior edge information and heightened accuracy. These results underscore the potential of our approach in advancing image processing techniques. [ABSTRACT FROM AUTHOR]

With the development of computer vision and digital image processing technology, image segmentation has become an important part of various image processing and image analysis. Since interactive segmentation can obtain more accurate results than automatic segmentation, the most representative Graph Cuts has gradually become a popular method in image segmentation. However, this algorithm has two significant disadvantages. On the one hand, if the background is complex or very similar to the foreground, the accuracy will be low; on the other hand, the algorithm is slow and the iteration process is complicated. To improve it, this paper proposes a new image segmentation algorithm based on quantum annealing and Graph Cuts. The algorithm beds the classical interactive image segmentation problem into a quantum optimization algorithm and obtains ideal image segmentation results on the D-Wave quantum annealer. Meanwhile, it is compared with the other three methods. Compared with MATLAB, the segmentation results are more beautiful, with an average precision higher than 5.27% and an average recall higher than 5.43%; the quantum annealing time is always lower than the simulated annealing time; and the success probability is more than twice that of the quantum approximate optimization algorithm. Therefore, it is concluded that this method is superior. [ABSTRACT FROM AUTHOR]