Search

The research groups in computer vision, graphics, and machine learning have dedicated a substantial amount of attention to the areas of 3D object reconstruction, augmentation, and registration. Deep learning is the predominant method used in artificial intelligence for addressing computer vision challenges. However, deep learning on three-dimensional data presents distinct obstacles and is now in its nascent phase. There have been significant advancements in deep learning specifically for three-dimensional data, offering a range of ways to address these issues. This study offers a comprehensive examination of the latest advancements in deep learning methodologies. We examine many benchmark models for the tasks of 3D object registration, augmentation, and reconstruction. We thoroughly analyse their architectures, advantages, and constraints. In summary, this report provides a comprehensive overview of recent advancements in three-dimensional deep learning and highlights unresolved research areas that will need to be addressed in the future. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT), projected to exceed 30 billion active device connections globally by 2025, presents an expansive attack surface. The frequent collection and dissemination of confidential data on these devices exposes them to significant security risks, including user information theft and denial-of-service attacks. This paper introduces a smart, network-based Intrusion Detection System (IDS) designed to protect IoT networks from distributed denial-of-service attacks. Our methodology involves generating synthetic images from flow-level traffic data of the Bot-IoT and the LATAM-DDoS-IoT datasets and conducting experiments within both supervised and self-supervised learning paradigms. Self-supervised learning is identified in the state of the art as a promising solution to replace the need for massive amounts of manually labeled data, as well as providing robust generalization. Our results showcase that self-supervised learning surpassed supervised learning in terms of classification performance for certain tests. Specifically, it exceeded the F1 score of supervised learning for attack detection by 4.83% and by 14.61% in accuracy for the multiclass task of protocol classification. Drawing from extensive ablation studies presented in our research, we recommend an optimal training framework for upcoming contrastive learning experiments that emphasize visual representations in the cybersecurity realm. This training approach has enabled us to highlight the broader applicability of self-supervised learning, which, in some instances, outperformed supervised learning transferability by over 5% in precision and nearly 1% in F1 score. [ABSTRACT FROM AUTHOR]

Unmanned aerial vehicle (UAV)-based object detection methods are widely used in traffic detection due to their high flexibility and extensive coverage. In recent years, with the increasing complexity of the urban road environment, UAV object detection algorithms based on deep learning have gradually become a research hotspot. However, how to further improve algorithmic efficiency in response to the numerous and rapidly changing road elements, and thus achieve high-speed and accurate road object detection, remains a challenging issue. Given this context, this paper proposes the high-efficiency multi-object detection algorithm for UAVs (HeMoDU). HeMoDU reconstructs a state-of-the-art, deep-learning-based object detection model and optimizes several aspects to improve computational efficiency and detection accuracy. To validate the performance of HeMoDU in urban road environments, this paper uses the public urban road datasets VisDrone2019 and UA-DETRAC for evaluation. The experimental results show that the HeMoDU model effectively improves the speed and accuracy of UAV object detection. [ABSTRACT FROM AUTHOR]

Scene text detection is an important research field in computer vision, playing a crucial role in various application scenarios. However, existing scene text detection methods often fail to achieve satisfactory results when faced with text instances of different sizes, shapes, and complex backgrounds. To address the challenge of detecting diverse texts in natural scenes, this paper proposes a multi-scale natural scene text detection method based on attention feature extraction and cascaded feature fusion. This method combines global and local attention through an improved attention feature fusion module (DSAF) to capture text features of different scales, enhancing the network's perception of text regions and improving its feature extraction capabilities. Simultaneously, an improved cascaded feature fusion module (PFFM) is used to fully integrate the extracted feature maps, expanding the receptive field of features and enriching the expressive ability of the feature maps. Finally, to address the cascaded feature maps, a lightweight subspace attention module (SAM) is introduced to partition the concatenated feature maps into several sub-space feature maps, facilitating spatial information interaction among features of different scales. In this paper, comparative experiments are conducted on the ICDAR2015, Total-Text, and MSRA-TD500 datasets, and comparisons are made with some existing scene text detection methods. The results show that the proposed method achieves good performance in terms of accuracy, recall, and F-score, thus verifying its effectiveness and practicality. [ABSTRACT FROM AUTHOR]

This document is an editorial for a special issue titled "Machine Learning in Computer Vision and Image Sensing: Theory and Applications." The editorial highlights the diverse applications of machine learning (ML) models in various domains such as medical imaging, signal processing, remote sensing, and human activity detection. The special issue includes 11 papers that cover topics such as image segmentation, fluvial navigation, Alzheimer's disease classification, pneumothorax detection, lung cancer malignancy prediction, amniotic fluid volume detection, COVID-19 detection, and Parkinson's disease detection. The papers showcase the progress and potential of ML models in computer vision applications and provide valuable insights for future research. [Extracted from the article]

Simple Summary: In the sorting process of Chinese softshell turtles, it is necessary to classify them based on their weight and accurately identify their plastron and carapace. This process requires heavy manual labor and complex mechanical processing methods. To improve processing efficiency and reduce costs, this article introduces machine vision technology, and a new image processing method is proposed. This method can estimate the weight of Chinese softshell turtles and accurately locate the positions of their plastron and carapace. The automation level of aquaculture can be greatly enhanced, and hardware costs can be reduced through software optimization through this approach. With the rapid development of the turtle breeding industry in China, the demand for automated turtle sorting is increasing. The automatic sorting of Chinese softshell turtles mainly consists of three parts: visual recognition, weight prediction, and individual sorting. This paper focuses on two aspects, i.e., visual recognition and weight prediction, and a novel method for the object detection and weight prediction of Chinese softshell turtles is proposed. In the individual sorting process, computer vision technology is used to estimate the weight of Chinese softshell turtles and classify them by weight. For the visual recognition of the body parts of Chinese softshell turtles, a color space model is proposed in this paper to separate the turtles from the background effectively. By applying multiple linear regression analysis for modeling, the relationship between the weight and morphological parameters of Chinese softshell turtles is obtained, which can be used to estimate the weight of turtles well. An improved deep learning object detection network is used to extract the features of the plastron and carapace of the Chinese softshell turtles, achieving excellent detection results. The mAP of the improved network reached 96.23%, which can meet the requirements for the accurate identification of the body parts of Chinese softshell turtles. [ABSTRACT FROM AUTHOR]

In reviewing this Special Issue, various topics have been addressed, predominantly machine learning techniques and heuristic search algorithms. Machine learning and computational intelligence are currently high-profile research areas attracting the attention of many researchers. In the first paper, L. Zhao and H. Jin improved the traditional vector-weighted optimization algorithm (INFO) and designed a promising optimization algorithm (IDEINFO) [[8]]. [Extracted from the article]

Despite the subjective and error-prone nature of manual visual inspection procedures, this type of inspection is still a common process in most construction projects. However, Automated Construction Inspection and Progress Monitoring (ACIPM) has the potential to improve inspection processes. The objective of this paper is to examine the applications, challenges, and future directions of ACIPM in a systematic review. It explores various application areas of ACIPM in two domains of (a) transportation construction inspection, and (b) building construction inspection. The review identifies key ACIPM tools and techniques including Laser Scanning (LS), Uncrewed Aerial Systems (UAS), Robots, Radio Frequency Identification (RFID), Augmented Reality (AR), Virtual Reality (VR), Computer Vision (CV), Deep Learning, and Building Information Modeling (BIM). It also explores the challenges in implementing ACIPM, including limited generalization, data quality and validity, data integration, and real-time considerations. Studying legal implications and ethical and social impacts are among the future directions in ACIPM that are pinpointed in this paper. As the main contribution, this paper provides a comprehensive understanding of ACIPM for academic researchers and industry professionals. [ABSTRACT FROM AUTHOR]

Street-view images can help us to better understand the city environment and its potential characteristics. With the development of computer vision and deep learning, the technology of semantic segmentation algorithms has become more mature. However, DeeplabV3+, which is commonly used in semantic segmentation, has shortcomings such as a large number of parameters, high requirements for computing resources, and easy loss of detailed information. Therefore, this paper proposes LM-DeeplabV3+, which aims to greatly reduce the parameters and computations of the model while ensuring segmentation accuracy. Firstly, the lightweight network MobileNetV2 is selected as the backbone network, and the ECA attention mechanism is introduced after MobileNetV2 extracts shallow features to improve the ability of feature representation; secondly, the ASPP module is improved, and on this basis, the EPSA attention mechanism is introduced to achieve cross-dimensional channel attention and important feature interaction; thirdly, a loss function named CL loss is designed to balance the training offset of multiple categories and better indicate the segmentation quality. This paper conducted experimental verification on the Cityspaces dataset, and the results showed that the mIoU reached 74.9%, which was an improvement of 3.56% compared to DeeplabV3+; and the mPA reached 83.01%, which was an improvement of 2.53% compared to DeeplabV3+. [ABSTRACT FROM AUTHOR]

Object detection is one of the fundamental tasks in computer vision, holding immense significance in the realm of intelligent mobile scenes. This paper proposes a hybrid cross-feature interaction (HCFI) attention module for object detection in intelligent mobile scenes. Firstly, the paper introduces multiple kernel (MK) spatial pyramid pooling (SPP) based on SPP and improves the channel attention using its structure. This results in a hybrid cross-channel interaction (HCCI) attention module with better cross-channel interaction performance. Additionally, we bolster spatial attention by incorporating dilated convolutions, leading to the creation of the cross-spatial interaction (CSI) attention module with superior cross-spatial interaction performance. By seamlessly combining the above two modules, we achieve an improved HCFI attention module without resorting to computationally expensive operations. Through a series of experiments involving various detectors and datasets, our proposed method consistently demonstrates superior performance. This results in a performance improvement of 1.53% for YOLOX on COCO and a performance boost of 2.05% for YOLOv5 on BDD100K. Furthermore, we propose a solution that combines HCCI and HCFI to address the challenge of extremely small output feature layers in detectors, such as SSD. The experimental results indicate that the proposed method significantly improves the attention capability of object detection in intelligent mobile scenes. [ABSTRACT FROM AUTHOR]

This study examined the development and trends in non-destructive testing technology for fruit quality. The status of the research field and the application hotspots were investigated to provide a reference for future research in this field. Relevant studies on the non-destructive testing of fruit quality published between 1993 and 2022 were identified in the core database Web of Science. The temporal distribution, spatial distribution, literature features, research progress, and leading research hotspots were quantified and visualised using bibliometrics. The findings revealed that there continues to be active research and publications on non-destructive testing technology for fruit quality, with a good development trend. China and the USA are the major contributors to research on non-destructive testing technology for fruit quality. The major research institutions include Zhejiang University and the United States Department of Agriculture. The major papers are published in Postharvest Biology and Technology and Acta Horticulturae, among others. These studies mainly focus on agriculture, food, and gardening, among other topics. The detection indices mainly concern internal quality, such as sugar degree and soluble solids, and apparent quality, such as hardness. The detection technologies mainly include electronic nose (E-nose) technology, machine vision technology, and spectral detection technology. In the future, technological developments in artificial intelligence and deep learning will further promote the maturation and application of non-destructive testing technologies for fruit quality. [ABSTRACT FROM AUTHOR]

Skin lesions are essential for the early detection and management of a number of dermatological disorders. Learning-based methods for skin lesion analysis have drawn much attention lately because of improvements in computer vision and machine learning techniques. A review of the most-recent methods for skin lesion classification, segmentation, and detection is presented in this survey paper. The significance of skin lesion analysis in healthcare and the difficulties of physical inspection are discussed in this survey paper. The review of state-of-the-art papers targeting skin lesion classification is then covered in depth with the goal of correctly identifying the type of skin lesion from dermoscopic, macroscopic, and other lesion image formats. The contribution and limitations of various techniques used in the selected study papers, including deep learning architectures and conventional machine learning methods, are examined. The survey then looks into study papers focused on skin lesion segmentation and detection techniques that aimed to identify the precise borders of skin lesions and classify them accordingly. These techniques make it easier to conduct subsequent analyses and allow for precise measurements and quantitative evaluations. The survey paper discusses well-known segmentation algorithms, including deep-learning-based, graph-based, and region-based ones. The difficulties, datasets, and evaluation metrics particular to skin lesion segmentation are also discussed. Throughout the survey, notable datasets, benchmark challenges, and evaluation metrics relevant to skin lesion analysis are highlighted, providing a comprehensive overview of the field. The paper concludes with a summary of the major trends, challenges, and potential future directions in skin lesion classification, segmentation, and detection, aiming to inspire further advancements in this critical domain of dermatological research. [ABSTRACT FROM AUTHOR]

Accelerometers are commonly used to measure vibrations for condition monitoring in mechanical and civil structures; however, their high cost and point-based measurement approach present practical limitations. With rapid advancements in computer vision and deep learning, research into tracking the motion of individual pixels with vision cameras has increased. The recently developed CoTracker, a transformer-based model, has demonstrated excellence in motion tracking, yet its performance in measuring structural vibrations has not been fully explored. This paper investigates the efficacy of the CoTracker model in extracting full-field structural vibrations using cameras. It is initially applied to capture the dense point movements in video sequences of a cantilever beam recorded using a high-speed camera. Subsequently, modal analysis using delay-embedding dynamic mode decomposition (DMD) is conducted to extract modal parameters including natural frequencies, damping ratios, and mode shapes. The results, benchmarked against those from a reference accelerometer and the Finite Element Method (FEM) result, demonstrate CoTracker's high potential for general applicability in structural vibration measurements. [ABSTRACT FROM AUTHOR]

Emotion recognition has become increasingly important in the field of Deep Learning (DL) and computer vision due to its broad applicability by using human–computer interaction (HCI) in areas such as psychology, healthcare, and entertainment. In this paper, we conduct a systematic review of facial and pose emotion recognition using DL and computer vision, analyzing and evaluating 77 papers from different sources under Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Our review covers several topics, including the scope and purpose of the studies, the methods employed, and the used datasets. The scope of this work is to conduct a systematic review of facial and pose emotion recognition using DL methods and computer vision. The studies were categorized based on a proposed taxonomy that describes the type of expressions used for emotion detection, the testing environment, the currently relevant DL methods, and the datasets used. The taxonomy of methods in our review includes Convolutional Neural Network (CNN), Faster Region-based Convolutional Neural Network (R-CNN), Vision Transformer (ViT), and "Other NNs", which are the most commonly used models in the analyzed studies, indicating their trendiness in the field. Hybrid and augmented models are not explicitly categorized within this taxonomy, but they are still important to the field. This review offers an understanding of state-of-the-art computer vision algorithms and datasets for emotion recognition through facial expressions and body poses, allowing researchers to understand its fundamental components and trends. [ABSTRACT FROM AUTHOR]

Pests have caused significant losses to agriculture, greatly increasing the detection of pests in the planting process and the cost of pest management in the early stages. At this time, advances in computer vision and deep learning for the detection of pests appearing in the crop open the door to the application of target detection algorithms that can greatly improve the efficiency of tomato pest detection and play an important technical role in the realization of the intelligent planting of tomatoes. However, in the natural environment, tomato leaf pests are small in size, large in similarity, and large in environmental variability, and this type of situation can lead to greater detection difficulty. Aiming at the above problems, a network target detection model based on deep learning, YOLONDD, is proposed in this paper. Designing a new loss function, NMIoU (Normalized Wasserstein Distance with Mean Pairwise Distance Intersection over Union), which improves the ability of anomaly processing, improves the model's ability to detect and identify objects of different scales, and improves the robustness to scale changes; Adding a Dynamic head (DyHead) with an attention mechanism will improve the detection ability of targets at different scales, reduce the number of computations and parameters, improve the accuracy of target detection, enhance the overall performance of the model, and accelerate the training process. Adding decoupled head to Head can effectively reduce the number of parameters and computational complexity and enhance the model's generalization ability and robustness. The experimental results show that the average accuracy of YOLONDD can reach 90.1%, which is 3.33% higher than the original YOLOv5 algorithm and is better than SSD, Faster R-CNN, YOLOv7, YOLOv8, RetinaNet, and other target detection networks, and it can be more efficiently and accurately utilized in tomato leaf pest detection. [ABSTRACT FROM AUTHOR]

Using instance segmentation and video inpainting provides a significant leap in real-time football video broadcast enhancements by removing potential visual distractions, such as an occasional person or another object accidentally occupying the frame. Despite its relevance and importance in the media industry, this area remains challenging and relatively understudied, thus offering potential for research. Specifically, the segmentation and inpainting of camera operator instances from video remains an underexplored research area. To address this challenge, this paper proposes a framework designed to accurately detect and remove camera operators while seamlessly hallucinating the background in real-time football broadcasts. The approach aims to enhance the quality of the broadcast by maintaining its consistency and level of engagement to retain and attract users during the game. To implement the inpainting task, firstly, the camera operators instance segmentation method should be developed. We used a YOLOv8 model for accurate real-time operator instance segmentation. The resulting model produces masked frames, which are used for further camera operator inpainting. Moreover, this paper presents an extensive "Cameramen Instances" dataset with more than 7500 samples, which serves as a solid foundation for future investigations in this area. The experimental results show that the YOLOv8 model performs better than other baseline algorithms in different scenarios. The precision of 95.5%, recall of 92.7%, mAP50-95 of 79.6, and a high FPS rate of 87 in low-volume environment prove the solution efficacy for real-time applications. [ABSTRACT FROM AUTHOR]

Photovoltaic (PV) fault detection is crucial because undetected PV faults can lead to significant energy losses, with some cases experiencing losses of up to 10%. The efficiency of PV systems depends upon the reliable detection and diagnosis of faults. The integration of Artificial Intelligence (AI) techniques has been a growing trend in addressing these issues. The goal of this systematic review is to offer a comprehensive overview of the recent advancements in AI-based methodologies for PV fault detection, consolidating the key findings from 31 research papers. An initial pool of 142 papers were identified, from which 31 were selected for in-depth review following the PRISMA guidelines. The title, objective, methods, and findings of each paper were analyzed, with a focus on machine learning (ML) and deep learning (DL) approaches. ML and DL are particularly suitable for PV fault detection because of their capacity to process and analyze large amounts of data to identify complex patterns and anomalies. This study identified several AI techniques used for fault detection in PV systems, ranging from classical ML methods like k-nearest neighbor (KNN) and random forest to more advanced deep learning models such as Convolutional Neural Networks (CNNs). Quantum circuits and infrared imagery were also explored as potential solutions. The analysis found that DL models, in general, outperformed traditional ML models in accuracy and efficiency. This study shows that AI methodologies have evolved and been increasingly applied in PV fault detection. The integration of AI in PV fault detection offers high accuracy and effectiveness. After reviewing these studies, we proposed an Artificial Neural Network (ANN)-based method for PV fault detection and classification. [ABSTRACT FROM AUTHOR]

The purpose of image dehazing is to remove the interference from weather factors in degraded images and enhance the clarity and color saturation of images to maximize the restoration of useful features. Single image dehazing is one of the most important tasks in the field of image restoration. In recent years, due to the progress of deep learning, single image dehazing has made great progress. With the success of Transformer in advanced computer vision tasks, some research studies also began to apply Transformer to image dehazing tasks and obtained surprising results. However, both the deconvolution-neural-network-based dehazing algorithm and Transformer based dehazing algorithm magnify their advantages and disadvantages separately. Therefore, this paper proposes a novel Transformer–Convolution fusion dehazing network (TCFDN), which uses Transformer's global modeling ability and convolutional neural network's local modeling ability to improve the dehazing ability. In the Transformer–Convolution fusion dehazing network, the classic self-encoder structure is used. This paper proposes a Transformer–Convolution hybrid layer, which uses an adaptive fusion strategy to make full use of the Swin-Transformer and convolutional neural network to extract and reconstruct image features. On the basis of previous research, this layer further improves the ability of the network to remove haze. A series of contrast experiments and ablation experiments not only proved that the Transformer–Convolution fusion dehazing network proposed in this paper exceeded the more advanced dehazing algorithm, but also provided solid and powerful evidence for the basic theory on which it depends. [ABSTRACT FROM AUTHOR]

Weak magnetic flux leak detection is one of the most important non-destructive testing and measurement methods for pipelines. Since different defects cause different damage, it is necessary to classify the different types of defects. Traditional machine learning methods of defect type identification mainly use feature analysis methods and rely on expert a priori knowledge and the ability of designers. These methods have the following weaknesses: a priori knowledge needs to be designed iteratively, and a priori knowledge design relies on expert experience. In recent years, the rapid development of deep learning methods in the field of machine vision has led to the development of defect analysis in the industry. However, most deep learning methods lack the ability to analyze both detailed information and the overall structure. In this paper, we propose graph convolution networks for splitting and fusing multiple graphs of detail graphs and a root graph. Detail information (detail graphs) provides detailed information for the detection of WMFLs. The structure information (root graph) provides structural information for the detection of WMFLs. This paper uses simulation data and experimental data to verify that the proposed method can identify stress defects and corrosion defects well. The paper explains the experimental results in detail to demonstrate the superiority of the method in industrial methods. [ABSTRACT FROM AUTHOR]

Simple Summary: Improving the accuracy of day-age detection of chickens is of great importance for chicken rearing. This paper focuses on the problem of classifying the age of chickens within 100 days. This paper proposes a high-precision federated learning-based model that can be applied to edge computing scenarios. Finally, our method can achieve an accuracy of 96.1%, which can fully meet the needs of application scenarios. Due to the booming development of computer vision technology and artificial intelligence algorithms, it has become more feasible to implement artificial rearing of animals in real production scenarios. Improving the accuracy of day-age detection of chickens is one of the examples and is of great importance for chicken rearing. This paper focuses on the problem of classifying the age of chickens within 100 days. Due to the huge amount of data and the different computing power of different devices in practical application scenarios, it is important to maximize the computing power of edge computing devices without sacrificing accuracy. This paper proposes a high-precision federated learning-based model that can be applied to edge computing scenarios. In order to accommodate different computing power in different scenarios, this paper proposes a dual-ended adaptive federated learning framework; in order to adapt to low computing power scenarios, this paper performs lightweighting operations on the mainstream model; and in order to verify the effectiveness of the model, this paper conducts a number of targeted experiments. Compared with AlexNet, VGG, ResNet and GoogLeNet, this model improves the classification accuracy to 96.1%, which is 14.4% better than the baseline model and improves the Recall and Precision by 14.8% and 14.2%, respectively. In addition, by lightening the network, our methods reduce the inference latency and transmission latency by 24.4 ms and 10.5 ms, respectively. Finally, this model is deployed in a real-world application and an application is developed based on the wechat SDK. [ABSTRACT FROM AUTHOR]

Shrimp farming has traditionally served as a crucial source of seafood and revenue for coastal countries. However, with the rapid development of society, conventional small-scale manual shrimp farming can no longer meet the increasing demand for rapid growth. As a result, it is imperative to continuously develop automation technology for efficient large-scale shrimp farming. Smart shrimp farming represents an innovative application of advanced technologies and management practices in shrimp aquaculture to expand the scale of production. Nonetheless, the use of these new technologies is not without difficulties, including the scarcity of public datasets and the high cost of labeling. In this paper, we focus on the application of advanced computer vision techniques to shrimp farming. To achieve this objective, we first establish a high-quality shrimp dataset for training various deep learning models. Subsequently, we propose a method that combines unsupervised learning with downstream instance segmentation tasks to mitigate reliance on large training datasets. Our experiments demonstrate that the method involving contrastive learning outperforms the direct fine-tuning of an instance segmentation model for shrimp in instance segmentation tasks. Furthermore, the concepts presented in this paper can extend to other fields that utilize computer vision technologies. [ABSTRACT FROM AUTHOR]

The sugar apple (Annona squamosa) is valued for its taste, nutritional richness, and versatility, making it suitable for fresh consumption and medicinal use with significant commercial potential. Widely found in the tropical Americas and Asia's tropical or subtropical regions, it faces challenges in post-harvest ripeness assessment, predominantly reliant on manual inspection, leading to inefficiency and high labor costs. This paper explores the application of computer vision techniques in detecting ripeness levels of harvested sugar apples and proposes an improved deep learning model (ECD-DeepLabv3+) specifically designed for ripeness detection tasks. Firstly, the proposed model adopts a lightweight backbone (MobileNetV2), reducing complexity while maintaining performance through MobileNetV2′s unique design. Secondly, it incorporates the efficient channel attention (ECA) module to enhance focus on the input image and capture crucial feature information. Additionally, a Dense ASPP module is introduced, which enhances the model's perceptual ability and expands the receptive field by stacking feature maps processed with different dilation rates. Lastly, the proposed model emphasizes the spatial information of sugar apples at different ripeness levels by the coordinate attention (CA) module. Model performance is validated using a self-made dataset of harvested optical images categorized into three ripeness levels. The proposed model (ECD-DeepLabv3+) achieves values of 89.95% for MIoU, 94.58% for MPA, 96.60% for PA, and 94.61% for MF1, respectively. Compared to the original DeepLabv3+, it greatly reduces the number of model parameters (Params) and floating-point operations (Flops) by 89.20% and 69.09%, respectively. Moreover, the proposed method could be directly applied to optical images obtained from the surface of the sugar apple, which provides a potential solution for the detection of post-harvest fruit ripeness. [ABSTRACT FROM AUTHOR]

Enhancing the environmental perception of autonomous vehicles (AVs) in intelligent transportation systems requires computer vision technology to be effective in detecting objects and obstacles, particularly in adverse weather conditions. Adverse weather circumstances present serious difficulties for object-detecting systems, which are essential to contemporary safety procedures, infrastructure for monitoring, and intelligent transportation. AVs primarily depend on image processing algorithms that utilize a wide range of onboard visual sensors for guidance and decisionmaking. Ensuring the consistent identification of critical elements such as vehicles, pedestrians, and road lanes, even in adverse weather, is a paramount objective. This paper not only provides a comprehensive review of the literature on object detection (OD) under adverse weather conditions but also delves into the ever-evolving realm of the architecture of AVs, challenges for automated vehicles in adverse weather, the basic structure of OD, and explores the landscape of traditional and deep learning (DL) approaches for OD within the realm of AVs. These approaches are essential for advancing the capabilities of AVs in recognizing and responding to objects in their surroundings. This paper further investigates previous research that has employed both traditional and DL methodologies for the detection of vehicles, pedestrians, and road lanes, effectively linking these approaches with the evolving field of AVs. Moreover, this paper offers an in-depth analysis of the datasets commonly employed in AV research, with a specific focus on the detection of key elements in various environmental conditions, and then summarizes the evaluation matrix. We expect that this review paper will help scholars to gain a better understanding of this area of research. [ABSTRACT FROM AUTHOR]

Street view imagery (SVI) is a rich source of information for architectural and urban analysis using computer vision techniques, but its integration with other building-level data sources requires an additional step of visual building identification. This step is particularly challenging in architecturally homogeneous, dense residential streets featuring narrow buildings, due to a combination of SVI geolocation errors and occlusions that significantly increase the risk of confusing a building with its neighboring buildings. This paper introduces a robust deep learning-based method to identify buildings across multiple street views taken at different angles and times, using global optimization to correct the position and orientation of street view panoramas relative to their surrounding building footprints. Evaluating the method on a dataset of 2000 street views shows that its identification accuracy (88%) outperforms previous deep learning-based methods (79%), while methods solely relying on geometric parameters correctly show the intended building less than 50% of the time. These results indicate that previous identification methods lack robustness to panorama pose errors when buildings are narrow, densely packed, and subject to occlusions, while collecting multiple views per building can be leveraged to increase the robustness of visual identification by ensuring that building views are consistent. [ABSTRACT FROM AUTHOR]

Recent technological developments have enabled computers to identify and categorize facial expressions to determine a person's emotional state in an image or a video. This process, called "Facial Expression Recognition (FER)", has become one of the most popular research areas in computer vision. In recent times, deep FER systems have primarily concentrated on addressing two significant challenges: the problem of overfitting due to limited training data availability, and the presence of expression-unrelated variations, including illumination, head pose, image resolution, and identity bias. In this paper, a comprehensive survey is provided on deep FER, encompassing algorithms and datasets that offer insights into these intrinsic problems. Initially, this paper presents a detailed timeline showcasing the evolution of methods and datasets in deep facial expression recognition (FER). This timeline illustrates the progression and development of the techniques and data resources used in FER. Then, a comprehensive review of FER methods is introduced, including the basic principles of FER (components such as preprocessing, feature extraction and classification, and methods, etc.) from the pro-deep learning era (traditional methods using handcrafted features, i.e., SVM and HOG, etc.) to the deep learning era. Moreover, a brief introduction is provided related to the benchmark datasets (there are two categories: controlled environments (lab) and uncontrolled environments (in the wild)) used to evaluate different FER methods and a comparison of different FER models. Existing deep neural networks and related training strategies designed for FER, based on static images and dynamic image sequences, are discussed. The remaining challenges and corresponding opportunities in FER and the future directions for designing robust deep FER systems are also pinpointed. [ABSTRACT FROM AUTHOR]

Video-based peripheral oxygen saturation (SpO2) estimation, utilizing solely RGB cameras, offers a non-contact approach to measuring blood oxygen levels. Previous studies set a stable and unchanging environment as the premise for non-contact blood oxygen estimation. Additionally, they utilized a small amount of labeled data for system training and learning. However, it is challenging to train optimal model parameters with a small dataset. The accuracy of blood oxygen detection is easily affected by ambient light and subject movement. To address these issues, this paper proposes a contrastive learning spatiotemporal attention network (CL-SPO2Net), an innovative semi-supervised network for video-based SpO2 estimation. Spatiotemporal similarities in remote photoplethysmography (rPPG) signals were found in video segments containing facial or hand regions. Subsequently, integrating deep neural networks with machine learning expertise enabled the estimation of SpO2. The method had good feasibility in the case of small-scale labeled datasets, with the mean absolute error between the camera and the reference pulse oximeter of 0.85% in the stable environment, 1.13% with lighting fluctuations, and 1.20% in the facial rotation situation. [ABSTRACT FROM AUTHOR]

Given the pronounced impact COVID-19 continues to have on society—infecting 700 million reported individuals and causing 6.96 million deaths—many deep learning works have recently focused on the virus's diagnosis. However, assessing severity has remained an open and challenging problem due to a lack of large datasets, the large dimensionality of images for which to find weights, and the compute limitations of modern graphics processing units (GPUs). In this paper, a new, iterative application of transfer learning is demonstrated on the understudied field of 3D CT scans for COVID-19 severity analysis. This methodology allows for enhanced performance on the MosMed Dataset, which is a small and challenging dataset containing 1130 images of patients for five levels of COVID-19 severity (Zero, Mild, Moderate, Severe, and Critical). Specifically, given the large dimensionality of the input images, we create several custom shallow convolutional neural network (CNN) architectures and iteratively refine and optimize them, paying attention to learning rates, layer types, normalization types, filter sizes, dropout values, and more. After a preliminary architecture design, the models are systematically trained on a simplified version of the dataset-building models for two-class, then three-class, then four-class, and finally five-class classification. The simplified problem structure allows the model to start learning preliminary features, which can then be further modified for more difficult classification tasks. Our final model CoSev boosts classification accuracies from below 60% at first to 81.57% with the optimizations, reaching similar performance to the state-of-the-art on the dataset, with much simpler setup procedures. In addition to COVID-19 severity diagnosis, the explored methodology can be applied to general image-based disease detection. Overall, this work highlights innovative methodologies that advance current computer vision practices for high-dimension, low-sample data as well as the practicality of data-driven machine learning and the importance of feature design for training, which can then be implemented for improvements in clinical practices. [ABSTRACT FROM AUTHOR]

This paper consists of a bibliometric study that covers the topic of 3D object detection from 2022 until the present day. It employs various analysis approaches that shed light on the leading authors, affiliations, and countries within this research domain alongside the main themes of interest related to it. The findings revealed that China is the leading country in this domain given the fact that it is responsible for most of the scientific literature as well as being a host for the most productive universities and authors in terms of the number of publications. China is also responsible for initiating a significant number of collaborations with various nations around the world. The most basic theme related to this field is deep learning, along with autonomous driving, point cloud, robotics, and LiDAR. The work also includes an in-depth review that underlines some of the latest frameworks that took on various challenges regarding this topic, the improvement of object detection from point clouds, and training end-to-end fusion methods using both camera and LiDAR sensors, to name a few. [ABSTRACT FROM AUTHOR]

The quality of cotton seed is of great significance to the production of cotton in the cotton industry. In order to reduce the workload of the manual sorting of cotton seeds and improve the quality of cotton seed sorting, this paper proposed an image-detection method of cotton seed damage based on an improved YOLOv5 algorithm. Images of cotton seeds with different degrees of damage were collected in the same environment. Cotton seeds of three different damage degrees, namely, undamaged, slightly damaged, and seriously damaged, were selected as the research objects. Labeling software was used to mark the images of these cotton seeds and the marked images were input into the improved YOLOv5s detection algorithm for appearance-based damage identification. The algorithm added the lightweight upsampling operator CARAFE to the original YOLOv5s detection algorithm and also improved the loss function. The experimental results showed that the mAP_0.5 value of the improved algorithm reached 99.5% and the recall rate reached 99.3% when the uncoated cotton seeds were detected. When detecting coated cotton seeds, the mAP_0.5 value of the improved algorithm reached 99.2% and the recall rate reached 98.9%. Compared with the traditional appearance-based damage detection approach, the improved YOLOv5s proposed in this paper improved the recognition accuracy and processing speed, and exhibited a better adaptability and generalization ability. Therefore, the proposed method can provide a reference for the appearance detection of crop seeds. [ABSTRACT FROM AUTHOR]

Waste disposal remains a challenge due to land availability, and environmental and health issues related to the main disposal method, landfilling. Combining computer vision (machine learning) and robotics to sort waste is a cost-effective solution for landfilling activities limitation. The objective of this study was to combine transfer and ensemble learning to process collected waste images and classify landfill waste into nine classes. Pretrained CNN models (Inception–ResNet-v2, EfficientNetb3, and DenseNet201) were used as base models to develop the ensemble network, and three other single CNN models (Models 1, 2, and 3). The single network performances were compared to the ensemble model. The waste dataset, initially grouped in two classes, was obtained from Kaggle, and reorganized into nine classes. Classes with a low number of data were improved by downloading additional images from Google search. The Ensemble Model showed the highest prediction precision (90%) compared to the precision of Models 1, 2, and 3, 86%, 87%, and 88%, respectively. All models had difficulties predicting overlapping classes, such as glass and plastics, and wood and paper/cardboard. The environmental costs for the Ensemble network, and Models 2 and 3, approximately 15 g CO2 equivalent per training, were lower than the 19.23 g CO2 equivalent per training for Model 1. [ABSTRACT FROM AUTHOR]

The main task in visual object tracking is to track a moving object in an image sequence. In this process, the object's trajectory and behavior can be described by calculating the object's position, velocity, acceleration, and other parameters or by memorizing the position of the object in each frame of the corresponding video. Therefore, visual object tracking can complete many more advanced tasks, has great performance in relation to real scenes, and is widely used in automated driving, traffic monitoring, human–computer interaction, and so on. Siamese-network-based trackers have been receiving a great deal of attention from the tracking community, but they have many drawbacks. This paper analyzes the shortcomings of the Siamese network tracker in detail, uses the method of feature multi-scale fusion to improve the Siamese network tracker, and proposes a new target-tracking framework to address its shortcomings. In this paper, a feature map with low-resolution but strong semantic information and a feature map with high-resolution and rich spatial information are integrated to improve the model's ability to depict an object, and the problem of scale change is solved by fusing features at different scales. Furthermore, we utilize the 3D Max Filtering module to suppress repeated predictions of features at different scales. Finally, our experiments conducted on the four tracking benchmarks OTB2015, VOT2016, VOT2018, and GOT10K show that the proposed algorithm effectively improves the tracking accuracy and robustness of the system. [ABSTRACT FROM AUTHOR]

The plum is a kind of delicious and common fruit with high edible value and nutritional value. The accurate and effective detection of plum fruit is the key to fruit number counting and pest and disease early warning. However, the actual plum orchard environment is complex, and the detection of plum fruits has many problems, such as leaf shading and fruit overlapping. The traditional method of manually estimating the number of fruits and the presence of pests and diseases used in the plum growing industry has disadvantages, such as low efficiency, a high cost, and low accuracy. To detect plum fruits quickly and accurately in a complex orchard environment, this paper proposes an efficient plum fruit detection model based on an improved You Only Look Once version 7(YOLOv7). First, different devices were used to capture high-resolution images of plum fruits growing under natural conditions in a plum orchard in Gulin County, Sichuan Province, and a dataset for plum fruit detection was formed after the manual screening, data enhancement, and annotation. Based on the dataset, this paper chose YOLOv7 as the base model, introduced the Convolutional Block Attention Module (CBAM) attention mechanism in YOLOv7, used Cross Stage Partial Spatial Pyramid Pooling–Fast (CSPSPPF) instead of Cross Stage Partial Spatial Pyramid Pooling(CSPSPP) in the network, and used bilinear interpolation to replace the nearest neighbor interpolation in the original network upsampling module to form the improved target detection algorithm YOLOv7-plum. The tested YOLOv7-plum model achieved an average precision (AP) value of 94.91%, which was a 2.03% improvement compared to the YOLOv7 model. In order to verify the effectiveness of the YOLOv7-plum algorithm, this paper evaluated the performance of the algorithm through ablation experiments, statistical analysis, etc. The experimental results showed that the method proposed in this study could better achieve plum fruit detection in complex backgrounds, which helped to promote the development of intelligent cultivation in the plum industry. [ABSTRACT FROM AUTHOR]

Deep learning models have revolutionized research fields like computer vision and natural language processing by outperforming traditional models in multiple tasks. However, the field of time series analysis, especially time series forecasting, has not seen a similar revolution, despite forecasting being one of the most prominent tasks of predictive data analytics. One crucial problem for time series forecasting is the lack of large, domain-independent benchmark datasets and a competitive research environment, e.g., annual large-scale challenges, that would spur the development of new models, as was the case for CV and NLP. Furthermore, the focus of time series forecasting research is primarily domain-driven, resulting in many highly individual and domain-specific datasets. Consequently, the progress in the entire field is slowed down due to a lack of comparability across models trained on a single benchmark dataset and on a variety of different forecasting challenges. In this paper, we first explore this problem in more detail and derive the need for a comprehensive, domain-unspecific overview of the state-of-the-art of commonly used datasets for prediction tasks. In doing so, we provide an overview of these datasets and improve comparability in time series forecasting by introducing a method to find similar datasets which can be utilized to test a newly developed model. Ultimately, our survey paves the way towards developing a single widely used and accepted benchmark dataset for time series data, built on the various frequently used datasets surveyed in this paper. [ABSTRACT FROM AUTHOR]

After the revival of deep learning in computer vision in 2012, SAR ship detection comes into the deep learning era too. The deep learning-based computer vision algorithms can work in an end-to-end pipeline, without the need of designing features manually, and they have amazing performance. As a result, it is also used to detect ships in SAR images. The beginning of this direction is the paper we published in 2017BIGSARDATA, in which the first dataset SSDD was used and shared with peers. Since then, lots of researchers focus their attention on this field. In this paper, we analyze the past, present, and future of the deep learning-based ship detection algorithms in SAR images. In the past section, we analyze the difference between traditional CFAR (constant false alarm rate) based and deep learning-based detectors through theory and experiment. The traditional method is unsupervised while the deep learning is strongly supervised, and their performance varies several times. In the present part, we analyze the 177 published papers about SAR ship detection. We highlight the dataset, algorithm, performance, deep learning framework, country, timeline, etc. After that, we introduce the use of single-stage, two-stage, anchor-free, train from scratch, oriented bounding box, multi-scale, and real-time detectors in detail in the 177 papers. The advantages and disadvantages of speed and accuracy are also analyzed. In the future part, we list the problem and direction of this field. We can find that, in the past five years, the AP50 has boosted from 78.8% in 2017 to 97.8 % in 2022 on SSDD. Additionally, we think that researchers should design algorithms according to the specific characteristics of SAR images. What we should do next is to bridge the gap between SAR ship detection and computer vision by merging the small datasets into a large one and formulating corresponding standards and benchmarks. We expect that this survey of 177 papers can make people better understand these algorithms and stimulate more research in this field. [ABSTRACT FROM AUTHOR]

Traditional ancient architecture is a symbolic product of cultural development and inheritance, with high social and cultural value. An automatic recognition model of ancient building types is one possible application of asymmetric systems, and it will be of great significance to be able to identify ancient building types via machine vision. In the context of Chinese traditional ancient buildings, this paper proposes a recognition method of ancient buildings, based on the improved asymmetric Inception V3 model. Firstly, the improved Inception V3 model adds a dropout layer between the global average pooling layer and the SoftMax classification layer to solve the overfitting problem caused by the small sample size of the ancient building data set. Secondly, migration learning and the ImageNet dataset are integrated into model training, which improves the speed of network training while solving the problems of the small scale of the ancient building dataset and insufficient model training. Thirdly, through ablation experiments, the effects of different data preprocessing methods and different dropout rates on the accuracy of model recognition were compared, to obtain the optimized model parameters. To verify the effectiveness of the model, this paper takes the ancient building dataset that was independently constructed by the South China University of Technology team as the experimental data and compares the recognition effect of the improved Inception V3 model proposed in this paper with several classical models. The experimental results show that when the data preprocessing method is based on filling and the dropout rate is 0.3, the recognition accuracy of the model is the highest; the accuracy rate of identifying ancient buildings using our proposed improved Inception V3 model can reach up to 98.64%. Compared with other classical models, the model accuracy rate has increased by 17.32%, and the average training time has accelerated by 2.29 times, reflecting the advantages of the model proposed in this paper. Finally, the improved Inception V3 model was loaded into the ancient building identification system to prove the practical application value of this research. [ABSTRACT FROM AUTHOR]

Human pose estimation refers to accurately estimating the position of the human body from a single RGB image and detecting the location of the body. It serves as the basis for several computer vision tasks, such as human tracking, 3D reconstruction, and autonomous driving. Improving the accuracy of pose estimation has significant implications for the advancement of computer vision. This paper addresses the limitations of single-branch networks in pose estimation. It presents a top-down single-target pose estimation approach based on multi-branch self-calibrating networks combined with graph convolutional neural networks. The study focuses on two aspects: human body detection and human body pose estimation. The human body detection is for athletes appearing in sports competitions, followed by human body pose estimation, which is divided into two methods: coordinate regression-based and heatmap test-based. To improve the accuracy of the heatmap test, the high-resolution feature map output from HRNet is used for deconvolution to improve the accuracy of single-target pose estimation recognition. [ABSTRACT FROM AUTHOR]

Visual tracking is a key research area in computer vision, as tracking technology is increasingly being applied in daily life, it has high-research significance. Visual tracking technology usually faces various challenging interference factors, among which, a similar background is one of the factors that has a greater impact on the tracking process. Kernelized Correlation Filter (KCF) tracking algorithm can track targets quickly by using circulant matrix, and has good tracking effect, so it is widely used in the tracking field. However, when the target is interfered by similar objects, the filter template in KCF cannot effectively distinguish between the target and the interfering object. This is because the filter only uses the texture gradient feature as the description object of the target, which will make the KCF algorithm extremely sensitive to the change of the target; therefore, the filter has difficultly making a judgment in the unstable scene, cannot accurately describe the target state, and finally leads to tracking failure. Therefore, this paper fuses Color Names (CN) on the basis of the original Histogram of Oriented Gradients (HOG) feature of KCF, which can obtain a more comprehensive feature representation, and realize the application of combined features to improve the anti-interference ability of KCF in complex scenes. In addition, this paper also uses the peak response of correlation filtering as the judgment condition to determine whether the current tracking result is stable. When the filter is in an unstable tracking state, the proposed algorithm will select the value with high confidence from its multiple responses as the candidate target of the Siamese network, and the deep learning network is used as the incremental learning method of the filter. The Channel Attention is introduced into the network layer, so that the network can adaptively reason and adjust the extracted universal features, and the enhanced feature information is used as the final discriminant basis. Finally, according to the response, the target with the smallest error compared with the target template is selected from multiple candidate targets as the final tracking result. The experimental results show that the average accuracy and average success rate of the proposed algorithm are significantly improved compared with the classical tracking algorithm, especially in dealing with similar target interference. [ABSTRACT FROM AUTHOR]

Simple Summary: The use of computer vision technology has improved the effectiveness of individual sheep identification, but existing methods face challenges such as large parameter sizes, slow recognition speeds, and difficult deployment. To address this issue, we have made improvements and optimizations based on the Retinaface face recognition model to create a balanced speed-accuracy sheep face recognition model. The optimized model has fewer parameters, simpler computations, faster inference speeds, and higher recognition accuracy, making it well-suited for deployment on resource-constrained edge devices. This research is expected to promote the application of deep learning-based sheep face recognition methods in production. The recognition of sheep faces based on computer vision has improved the efficiency and effectiveness of individual sheep identification, providing technical support for the development of smart farming. However, current recognition models have problems such as large parameter sizes, slow recognition speed, and difficult deployment. Therefore, this paper proposes an efficient and fast basic module called Eblock and uses it to build a lightweight sheep face recognition model called SheepFaceNet, which achieves the best balance between speed and accuracy. SheepFaceNet includes two modules: SheepFaceNetDet for detection and SheepFaceNetRec for recognition. SheepFaceNetDet uses Eblock to construct the backbone network to enhance feature extraction capability and efficiency, designs a bidirectional FPN layer (BiFPN) to enhance geometric location ability, and optimizes the network structure, which affects inference speed, to achieve fast and accurate sheep face detection. SheepFaceNetRec uses Eblock to construct the feature extraction network, uses ECA channel attention to improve the effectiveness of feature extraction, and uses multi-scale feature fusion to achieve fast and accurate sheep face recognition. On our self-built sheep face dataset, SheepFaceNet recognized 387 sheep face images per second with an accuracy rate of 97.75%, achieving an advanced balance between speed and accuracy. This research is expected to further promote the application of deep-learning-based sheep face recognition methods in production. [ABSTRACT FROM AUTHOR]

Human activity recognition (HAR) is an important research area with a wide range of application scenarios, such as smart homes, healthcare, abnormal behavior detection, etc. Wearable sensors, computer vision, radar, and other technologies are commonly used to detect human activity. However, they are severely limited by issues such as cost, lighting, context, and privacy. Therefore, this paper explores a high-performance method of using channel state information (CSI) to identify human activities, which is a deep learning-based spatial module-temporal convolutional network (SM-TCNNET) model. The model consists of a spatial feature extraction module and a temporal convolutional network (TCN) that can extract the spatiotemporal features in CSI signals well. In this paper, extensive experiments are conducted on the self-picked dataset and the public dataset (StanWiFi), and the results show that the accuracy reaches 99.93% and 99.80%, respectively. Compared with the existing methods, the recognition accuracy of the SM-TCNNET model proposed in this paper is improved by 1.8%. [ABSTRACT FROM AUTHOR]

Gait recognition, also known as walking pattern recognition, has expressed deep interest in the computer vision and biometrics community due to its potential to identify individuals from a distance. It has attracted increasing attention due to its potential applications and non-invasive nature. Since 2014, deep learning approaches have shown promising results in gait recognition by automatically extracting features. However, recognizing gait accurately is challenging due to the covariate factors, complexity and variability of environments, and human body representations. This paper provides a comprehensive overview of the advancements made in this field along with the challenges and limitations associated with deep learning methods. For that, it initially examines the various gait datasets used in the literature review and analyzes the performance of state-of-the-art techniques. After that, a taxonomy of deep learning methods is presented to characterize and organize the research landscape in this field. Furthermore, the taxonomy highlights the basic limitations of deep learning methods in the context of gait recognition. The paper is concluded by focusing on the present challenges and suggesting several research directions to improve the performance of gait recognition in the future. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of constructing a real-time 2D map for driving scenes from a single monocular RGB image. We presented a method based on three neural networks (depth estimation, 3D object detection, and semantic segmentation). We proposed a depth estimation neural network architecture that is fast and accurate in comparison with the state-of-the-art models. We designed our model to work in real time on light devices (such as an NVIDIA Jetson Nano and smartphones). The model is based on an encoder–decoder architecture with complex loss functions, i.e., normal loss, VNL, gradient loss (dx, dy), and mean absolute error. Our results show competitive results in comparison with the state-of-the-art methods, as our method is 30 times faster and smaller. [ABSTRACT FROM AUTHOR]

Identifying individual pigs is crucial for efficient breeding, health management, and disease control in modern farming. Traditional animal face identification methods are labor-intensive and prone to inaccuracies, while existing CNN-based pig face recognition models often struggle with high computational demands, large sizes, and reliance on extensive labeled data, which limit their practical application. This paper addresses these challenges by proposing a novel, decoupled approach to pig face recognition that separates detection from identification. This strategy employs a detection model as a pre-processing step, significantly reducing the need for extensive re-annotation for new datasets. Additionally, the paper introduces a method that integrates offline knowledge distillation with a lightweight pig face recognition model, aiming to build an efficient and embedding-friendly system. To achieve these objectives, the study constructs a small-scale, high-quality pig face detection dataset consisting of 1500 annotated images from a selection of 20 pigs. An independent detection model, trained on this dataset, then autonomously generates a large-scale pig face recognition dataset with 56 pig classes. In the face recognition stage, a robust teacher model guides the student model through a distillation process informed by a knowledge distillation loss, enabling the student model to learn relational features from the teacher. Experimental results confirm the high accuracy of the pig face detection model on the small-scale detection dataset and the ability to generate a large-scale dataset for pig face recognition on unlabeled data. The recognition experiments further verify that the distilled lightweight model outperforms its non-distilled counterparts and approaches the performance of the teacher model. This scalable, cost-effective solution shows significant promise for broader computer vision applications beyond agriculture. [ABSTRACT FROM AUTHOR]

In complex citrus orchard environments, light changes, branch shading, and fruit overlapping impact citrus detection accuracy. This paper proposes the citrus detection model YOLO-DCA in complex environments based on the YOLOv7-tiny model. We used depth-separable convolution (DWConv) to replace the ordinary convolution in ELAN, which reduces the number of parameters of the model; we embedded coordinate attention (CA) into the convolution to make it a coordinate attention convolution (CAConv) to replace the ordinary convolution of the neck network convolution; and we used a dynamic detection head to replace the original detection head. We trained and evaluated the test model using a homemade citrus dataset. The model size is 4.5 MB, the number of parameters is 2.1 M, mAP is 96.98%, and the detection time of a single image is 5.9 ms, which is higher than in similar models. In the application test, it has a better detection effect on citrus in occlusion, light transformation, and motion change scenes. The model has the advantages of high detection accuracy, small model space occupation, easy application deployment, and strong robustness, which can help citrus-picking robots and improve their intelligence level. [ABSTRACT FROM AUTHOR]

The species and population size of marine fish are important for maintaining the ecological environment and reflecting climate change. Traditional fish detection methods mainly rely on manual or traditional computer vision, which has disadvantages such as complex design, low detection accuracy, and poor generalization. The widespread use of ocean observation systems has accumulated a large number of images and videos, which makes the application of deep learning on marine fish detection possible. In this paper, a real-time high-precision fish detection algorithm based on YOLOv5s is constructed. Considering the enhancement of the channel representation and spatial interaction ability of the model, the attention mechanism and gated convolution are introduced, respectively, and GhostNet is introduced to lighten the model. Through a series of model comparisons, two improved models, S-Head-Ghost-Fish9 and S-SE-HorBlock-Head-Ghost-Fish9, are finally obtained. Compared with the original model, in terms of model size, the former reduces by 19% and the latter increases by 9.5%; in terms of computation, the former reduces by 15.7% and the latter reduces by 3.1%; in terms of detection speed, both take about 17 ms to detect a single image, and both can meet the real-time detection requirements; in terms of detection accuracy, the former improves by 3% and the latter by 3.6%. Compared with the latest detection algorithms of YOLOv6 and YOLOv8, the detection accuracy is slightly lower than 1%, but the model size and computation amount are only 1/3 to 1/2 of them. The improved models can help assess the population size and growth of the fish, which is of great significance in maintaining the stability of the fish population. [ABSTRACT FROM AUTHOR]

As urbanization increases, streetlights have become significant consumers of electrical power, making it imperative to develop effective control methods for sustainability. This paper offers a comprehensive review on control methods of smart streetlight systems, setting itself apart by introducing a novel light scheme framework that provides a structured classification of various light control patterns, thus filling an existing gap in the literature. Unlike previous studies, this work dives into the technical specifics of individual research papers and methodologies, ranging from basic to advanced control methods like computer vision and deep learning, while also assessing the energy consumption associated with each approach. Additionally, the paper expands the discussion to explore alternative functionalities for streetlights, such as serving as communication networks, environmental monitors, and electric vehicle charging stations. This multidisciplinary research aims to be a pivotal resource for both academics and industry professionals, laying the groundwork for future innovation and sustainable solutions in urban lighting. [ABSTRACT FROM AUTHOR]

This paper presents a novel approach to training a real-world object detection system based on synthetic data utilizing state-of-the-art technologies. Training an object detection system can be challenging and time-consuming as machine learning requires substantial volumes of training data with associated metadata. Synthetic data can solve this by providing unlimited desired training data with automatic generation. However, the main challenge is creating a balanced dataset that closes the reality gap and generalizes well when deployed in the real world. A state-of-the-art game engine, Unreal Engine 4, was used to approach the challenge of generating a photorealistic dataset for deep learning model training. In addition, a comprehensive domain randomized environment was implemented to create a robust dataset that generalizes the training data well. The randomized environment was reinforced by adding high-dynamic-range image scenes. Finally, a modern neural network was used to train the object detection system, providing a robust framework for an adaptive and self-learning model. The final models were deployed in simulation and in the real world to evaluate the training. The results of this study show that it is possible to train a real-world object detection system on synthetic data. However, the models showcase a lot of potential for improvements regarding the stability and confidence of the inference results. In addition, the paper provides valuable insight into how the number of assets and training data influence the resulting model. [ABSTRACT FROM AUTHOR]

This paper focuses on different types of backbone networks for machine learning architectures which perform classification of knee Magnetic Resonance Imaging (MRI) images. This paper aims to compare different types of feature extraction networks for the same classification task, in terms of accuracy and performance. Multiple variations of machine learning models were trained based on the MRNet architecture, choosing AlexNet, ResNet, VGG-11, VGG-16, and Efficientnet as the backbone. The models were evaluated on the MRNet validation dataset, computing Area Under the Receiver Operating Characteristics Curve (ROC-AUC), accuracy, f1 score, and Cohen's Kappa as evaluation metrics. The MRNet-VGG16 model variant shows the best results for Anterior Cruciate Ligament (ACL) tear detection. For general abnormality detection, MRNet-VGG16 is dominated by MRNet-Resnet in confidence between 0.5 and 0.75 and by MRNet-VGG11 for confidence more than 0.8. Due to the non-uniform nature of backbone network performance on different MRI planes, it is advisable to use an LR ensemble of: VGG16 on a coronal plane for all classification tasks; on an axial plane for abnormality and ACL tear detection; Alexnet on a sagittal plane for abnormality detection, and an axial plane for meniscal tear detection; and VGG11 on a sagittal plane for ACL tear detection. The results also indicate that the Cohen's Kappa metric is valuable in model evaluation for the MRNet dataset, as it provides deeper insights on classification decisions. [ABSTRACT FROM AUTHOR]

This paper introduces the Color Histogram Contouring (CHC) method, a new training-less approach to object detection that emphasizes the distinctive features in chrominance components. By building a chrominance-rich feature vector with a bin size of 1, the proposed CHC method exploits the precise information in chrominance features without increasing bin sizes, which can lead to false detections. This feature vector demonstrates invariance to lighting changes and is designed to mimic the opponent color axes used by the human visual system. The proposed CHC algorithm iterates over non-zero histogram bins of unique color features in the model, creating a feature vector for each, and emphasizes those matching in both the scene and model histograms. When both model and scene histograms for these unique features align, it ensures the presence of the model in the scene image. Extensive experiments across various scenarios show that the proposed CHC technique outperforms the benchmark training-less Swain and Ballard method and the algorithm of Viola and Jones. Additionally, a comparative experiment with the state-of-the-art You Only Look Once (YOLO) technique reveals that the proposed CHC technique surpasses YOLO in scenarios with limited training data, highlighting a significant advancement in training-less object detection. This approach offers a valuable addition to computer vision, providing an effective training-less solution for real-time autonomous robot localization and mapping in unknown environments. [ABSTRACT FROM AUTHOR]

Human–object interaction (HOI) detection identifies a "set of interactions" in an image involving the recognition of interacting instances and the classification of interaction categories. The complexity and variety of image content make this task challenging. Recently, the Transformer has been applied in computer vision and received attention in the HOI detection task. Therefore, this paper proposes a novel Part Refinement Tandem Transformer (PRTT) for HOI detection. Unlike the previous Transformer-based HOI method, PRTT utilizes multiple decoders to split and process rich elements of HOI prediction and introduces a new part state feature extraction (PSFE) module to help improve the final interaction category classification. We adopt a novel prior feature integrated cross-attention (PFIC) to utilize the fine-grained partial state semantic and appearance feature output obtained by the PSFE module to guide queries. We validate our method on two public datasets, V-COCO and HICO-DET. Compared to state-of-the-art models, the performance of detecting human–object interaction is significantly improved by the PRTT. [ABSTRACT FROM AUTHOR]

Point cloud registration is a fundamental task in computer vision and graphics, which is widely used in 3D reconstruction, object tracking, and atlas reconstruction. Learning-based optimization and deep learning methods have been widely developed in pairwise registration due to their own distinctive advantages. Deep learning methods offer greater flexibility and enable registering unseen point clouds that are not trained. Learning-based optimization methods exhibit enhanced robustness and stability when handling registration under various perturbations, such as noise, outliers, and occlusions. To leverage the strengths of both approaches to achieve a less time-consuming, robust, and stable registration for multiple instances, we propose a novel computational framework called SGRTmreg for multiple pairwise registrations in this paper. The SGRTmreg framework utilizes three components—a Searching scheme, a learning-based optimization method called Graph-based Reweighted discriminative optimization (GRDO), and a Transfer module to achieve multi-instance point cloud registration.Given a collection of instances to be matched, a template as a target point cloud, and an instance as a source point cloud, the searching scheme selects one point cloud from the collection that closely resembles the source. GRDO then learns a sequence of regressors by aligning the source to the target, while the transfer module stores and applies the learned regressors to align the selected point cloud to the target and estimate the transformation of the selected point cloud. In short, SGRTmreg harnesses a shared sequence of regressors to register multiple point clouds to a target point cloud. We conduct extensive registration experiments on various datasets to evaluate the proposed framework. The experimental results demonstrate that SGRTmreg achieves multiple pairwise registrations with higher accuracy, robustness, and stability than the state-of-the-art deep learning and traditional registration methods. [ABSTRACT FROM AUTHOR]