Computer Vision Research Articles

A systematic review on feature extraction methods and deep learning models for detection of cancerous lung nodules at an early stage -The Recent trends and challenges.

Lung cancer is one of the most common life-threatening worldwide cancers affecting both the male and female populations. The appearance of nodules in the scan image is an early indication of the development of cancer cells in the lung. The Low Dose Computed Tomography screening technique is used for the early detection of cancer nodules. Therefore, with more Computed Tomography (CT) lung profiles, an automated lung nodule analysis system can be utilized through image processing techniques and neural network algorithms. A CT image of the lung consists of many elements such as blood vessels, ribs, nodules, sternum, bronchi, and nodules. These nodules can be both benign and malignant, where the latter leads to lung cancer. Detecting them at an earlier stage can increase life expectancy by up to 5 to 10 years. To analyze only the nodules from the profile, respected features are calculated using image processing techniques. Based on the review, textural features were promising ones in medical image analysis and for solving computer vision problems. The significance of extracting the hidden features (textural) enables the higher performance of deep learning algorithms (DL) predominantly in medical imaging, which has resulted in improvement in accuracy in recent years. Earlier detection of cancerous lung nodules is possible through the combination of multi-featured extraction and classification techniques using image data. In this paper, we discussed the overview of lung cancer along with publicly available datasets for research purposes. The primary objective of the paper is to provide the importance of textural features when combined with different deep-learning models. It gives insights into their advantages, disadvantages, and limitations regarding possible research gaps. The paper compared the recent studies of deep learning models with and without feature extraction and concluded that DL models that include feature extraction were better than the others.

Image–Text Matching Model Based on CLIP Bimodal Encoding

Image–text matching is a fundamental task in the multimodal research field, connecting computer vision and natural language processing by aligning visual content with corresponding textual descriptions. Accurate matching is critical for applications such as image captioning and text-based image retrieval yet remains challenging due to the differences in data modalities. This paper addresses these challenges by proposing a robust image–text matching model inspired by Contrastive Language–Image Pre-training (CLIP). Our approach employs the Vision Transformer (ViT) model as the image encoder and Bidirectional Encoder Representations from Transformers (Bert) as the text encoder, integrating these into a shared vector space to measure semantic similarity. We enhance the model’s training efficiency using the LiT-tuning paradigm to optimize learning through a cosine decay strategy for dynamic adjustment of the learning rate. We validate our method on two benchmark datasets, WuKong and Flickr30k, demonstrating that our model achieves superior performance and significantly improves key evaluation metrics. The results underscore the model’s effectiveness in achieving accurate and robust image–text alignment.

Robotic Grasping Detection Algorithm Based on 3D Vision Dual-Stream Encoding Strategy

The automatic generation of stable robotic grasping postures is crucial for the application of computer vision algorithms in real-world settings. This task becomes especially challenging in complex environments, where accurately identifying the geometric shapes and spatial relationships between objects is essential. To enhance the capture of object pose information in 3D visual scenes, we propose a planar robotic grasping detection algorithm named SU-Grasp, which simultaneously focuses on local regions and long-distance relationships. Built upon a U-shaped network, SU-Grasp introduces a novel dual-stream encoding strategy using the Swin Transformer combined with spatial semantic enhancement. Compared to existing baseline methods, our algorithm achieves superior performance across public datasets, simulation tests, and real-world scenarios, highlighting its robust understanding of complex spatial environments.

Realizing Bidirectional Photocurrent in Monolithic Dual‐Mode Device for Neuromorphic Vision and Logically‐Encrypted Transmission

AbstractDue to significant response contradictions, unidirectional carrier transmission of typical semiconductor p‐n junctions limits integrated sensors and artificial synapses in a monolithic device. In this work, a bipolar p‐n junction designed by employing the hydrogel/p‐GaN local contact interface with p‐GaN/(In,Ga)N heterojunction, demonstrating the bidirectional photocurrent and sensor/artificial synapse dual‐mode device successfully. After modifying Au nanoparticles, the negative (positive) photocurrent is increased by 900% (300%) under 365 nm (520 nm) light to achieve a more balanced bipolar carrier dynamic. Such a regulation of photocurrent is found to be attributed to the promotion of hydrogen evolution reaction (HER) at the hydrogel/Au/p‐GaN interface. Thus, the device achieves the ultrahigh paired‐pulse facilitation (PPF) index of 243% under self‐powered condition. Moreover, the implementation of plasticity from short‐term to long‐term demonstrates its adaptability in neural morphology visual recognition. Beyond independent working mode, the function of five classic logic gates is achieved under three‐terminal operation. Finally, an encrypted wireless communication system using dual‐channel light signals demonstrates the extensive application potential of the monolithic device. Therefore, this work presents a viable construction blueprint to meet the demands of next‐generation all‐in‐one optoelectronics for intricate application scenarios.

On Efficient Training of Large-Scale Deep Learning Models

The field of deep learning has witnessed significant progress in recent times, particularly in areas such as computer vision (CV), natural language processing (NLP), and speech. The use of large-scale models trained on vast amounts of data holds immense promise for practical applications, enhancing industrial productivity and facilitating social development. However, it suffers extremely from the unstable training process and stringent requirements of computational resources. With the increasing demands on the adaption of computational capacity, though numerous studies have explored the efficient training field to a certain extent, a comprehensive summarization/guideline on those general acceleration techniques of training large-scale deep learning models is still much anticipated. In this survey, we present a detailed review of the general techniques for training acceleration. We consider the fundamental update formulation and split its basic components into five main perspectives: (1) “data-centric,” including dataset regularization, data sampling, and data-centric curriculum learning techniques, which can significantly reduce the computational complexity of the data samples; (2) “model-centric,” including acceleration of basic modules, compression training, model initialization, and model-centric curriculum learning techniques, which focus on accelerating the training via reducing the calculations on parameters and providing better initialization; (3) “optimization-centric,” including the selection of learning rate, the employment of large batch size, the designs of efficient objectives, and model average techniques, which pay attention to the training policy and improving the generality for the large-scale models; (4) “budgeted training,” including some distinctive acceleration methods on source-constrained situations, e.g., for limitation on the total iterations; and (5) “system-centric,” including some efficient distributed frameworks and open source libraries that provide adequate hardware support for the implementation of the above-mentioned acceleration algorithms. By presenting this comprehensive taxonomy, our survey presents a comprehensive review to understand the general mechanisms within each component and their joint interaction. Meanwhile, we further provide a detailed analysis and discussion of future works on the development of general acceleration techniques, which could inspire us to re-think and design novel efficient paradigms. Overall, we hope that this survey will serve as a valuable guideline for general efficient training.

Comparison of CNN-Based Architectures for Detection of Different Object Classes

(1) Background: Detecting people and technical objects in various situations, such as natural disasters and warfare, is critical to search and rescue operations and the safety of civilians. A fast and accurate detection of people and equipment can significantly increase the effectiveness of search and rescue missions and provide timely assistance to people. Computer vision and deep learning technologies play a key role in detecting the required objects due to their ability to analyze big volumes of visual data in real-time. (2) Methods: The performance of the neural networks such as You Only Look Once (YOLO) v4-v8, Faster R-CNN, Single Shot MultiBox Detector (SSD), and EfficientDet has been analyzed using COCO2017, SARD, SeaDronesSee, and VisDrone2019 datasets. The main metrics for comparison were mAP, Precision, Recall, F1-Score, and the ability of the neural network to work in real-time. (3) Results: The most important metrics for evaluating the efficiency and performance of models for a given task are accuracy (mAP), F1-Score, and processing speed (FPS). These metrics allow us to evaluate both the accuracy of object recognition and the ability to use the models in real-world environments where high processing speed is important. (4) Conclusion: Although different neural networks perform better on certain types of metrics, YOLO outperforms them on all metrics, showing the best results of mAP-0.88, F1-0.88, and FPS-48, so the focus was on these models.

Enhancing YOLOv8’s Performance in Complex Traffic Scenarios: Optimization Design for Handling Long-Distance Dependencies and Complex Feature Relationships

In recent years, the field of deep learning and computer vision has increasingly focused on the problem of vehicle target detection, becoming the forefront of many technological innovations. YOLOv8, as an efficient vehicle target detection model, has achieved good results in many scenarios. However, when faced with complex traffic scenarios, such as occluded targets, small target detection, changes in lighting, and variable weather conditions, YOLOv8 still has insufficient detection accuracy and robustness. To address these issues, this paper delves into the optimization strategies of YOLOv8 in the field of vehicle target detection, focusing on the EMA module in the backbone part and replacing the original SPPF module with focal modulation technology, all of which effectively improved the model’s performance. At the same time, modifications to the head part were approached with caution to avoid unnecessary interference with the original design. The experiment used the UA-DETRAC dataset, which contains a variety of traffic scenarios, a rich variety of vehicle types, and complex dynamic environments, making it suitable for evaluating and validating the performance of traffic monitoring systems. The 5-fold cross-validation method was used to ensure the reliability and comprehensiveness of the evaluation results. The final results showed that the improved model’s precision rate increased from 0.859 to 0.961, the recall rate from 0.83 to 0.908, and the mAP50 from 0.881 to 0.962. Meanwhile, the optimized YOLOv8 model demonstrated strong robustness in terms of detection accuracy and the ability to adapt to complex environments.

Harnessing Artificial Intelligence for Wildlife Conservation

The rapid decline in global biodiversity demands innovative conservation strategies. This paper examines the use of artificial intelligence (AI) in wildlife conservation, focusing on the Conservation AI platform. Leveraging machine learning and computer vision, Conservation AI detects and classifies animals, humans, and poaching-related objects using visual spectrum and thermal infrared cameras. The platform processes these data with convolutional neural networks (CNNs) and transformer architectures to monitor species, including those that are critically endangered. Real-time detection provides the immediate responses required for time-critical situations (e.g., poaching), while non-real-time analysis supports long-term wildlife monitoring and habitat health assessment. Case studies from Europe, North America, Africa, and Southeast Asia highlight the platform’s success in species identification, biodiversity monitoring, and poaching prevention. The paper also discusses challenges related to data quality, model accuracy, and logistical constraints while outlining future directions involving technological advancements, expansion into new geographical regions, and deeper collaboration with local communities and policymakers. Conservation AI represents a significant step forward in addressing the urgent challenges of wildlife conservation, offering a scalable and adaptable solution that can be implemented globally.

Television and the Origins of Visual Pattern Recognition AI in Japan: a ‘Quest for a Seeing Machine’

Abstract This article traces the genealogy of contemporary data-driven computer vision to developments at the Japan Broadcasting Corporation (NHK) in the 1960s. Specifically, it examines NHK’s Visual and Auditory Information Science Unit and its role in the invention of the world’s first deep convolutional neural network. The use of television to collect viewer behaviour data enabled modelling of eye-brain information processing, in particular mechanisms of feature extraction. This, in turn, linked Fukushima Kunihiko’s formative work on signal compression to the development of a pattern recognition machine, resulting in the creation of the world’s first convolutional neural network. Recovering this history is important for two reasons. First, it helps counter a trend of ‘digital universalism’ that covertly homogenizes local differences into a single culture of artificial intelligence in the Cold-War-era US. Second, it reveals the largely ignored role of television in the genesis of digital image technologies and AI more broadly.

Person in Uniforms Re-Identification

Person in Uniforms Re-identification (PU-ReID) is an emerging computer vision task for various intelligent video surveillance applications. PU-ReID is much understudied due to the absence of large-scale annotated datasets, also this task is extremely challenging because many individuals captured in surveillance videos wear same clothing, introducing significant interference for retrieval tasks owing to the high visual similarity of outfits and subtle differences among individuals. This research initiates the exploration of person in uniform re-identification, a novel and challenging task tailored for real industrial scenarios. To address these issues, a novel framework is proposed for PU-ReID, which aims to reduce the visual impact of similar uniforms and learn the unique cues derived from human parts and detailed visual features. Specifically, several novel techniques are built in this study: first, a uniform feature separation method with orthogonal constraints is proposed to extract non-uniform features. Second, multi-view subspace feature alignment is introduced to integrate soft-biometrics including optics-related visual features, contextual information of human parts, and cloth-invariant biometric features. In addition, to close the gap between academic research and real world settings, a new person in uniforms ReID dataset named PU-151 is constructed, which consists of 151 gas station employees in uniforms from 1,488 videos. At last, extensive experiments conducted on five datasets demonstrate that the proposed approach significantly outperforms the state-of-the-art methods. This advancement can drive further developments in re-identification and person search technologies.

SSDSNet: A Dual-Stream State-Space Network for Efficient Image Denoising with Long-Range Dependency Modeling

Abstract In recent years, significant progress has been made in image denoising, driven by advancements in modern deep neural networks such as Convolutional Neural Networks
(CNNs) and Transformers. However, existing denoising models often face limitations due to inherent local inductive biases or the quadratic computational complexity of their architectures. Recently, Selective Structured State Space Models, exemplified by Mamba, have shown considerable potential in modeling long-range dependencies with linear complexity. Despite this promise, their application in low-level computer vision tasks remains underexplored. In this study, we introduce a straightforward yet effective model, the Dual-Stream State-Space Module (DSSM), specifically designed for image denoising. The DSSM serves as the core component of our model, combining convolutional operations with the
Mamba model to enhance the capability of capturing both local features and global context.
This dual-stream approach leverages the inherent structure of local image patches while effectively modeling long-range dependencies, resulting in feature representations tailored for denoising tasks. Extensive experiments demonstrate the effectiveness of our method. For instance, on the SIDD dataset, our model achieves performance on par with state-of-the-art methods, maintaining a global receptive field while operating with similar computational costs.

TouchView: Mid-Air Touch on Zoomable 2D View for Distant Freehand Selection on a Virtual Reality User Interface

Selection is a fundamental interaction element in virtual reality (VR) and 3D user interfaces (UIs). Raycasting, one of the most common object selection techniques, is known to have difficulties in selecting small or distant objects. Meanwhile, recent advancements in computer vision technology have enabled seamless vision-based hand tracking in consumer VR headsets, enhancing accessibility to freehand mid-air interaction and highlighting the need for further research in this area. This study proposes a new technique called TouchView, which utilizes a virtual panel with a modern adaptation of the Through-the-Lens metaphor to improve freehand selection for VR UIs. TouchView enables faster and less demanding target selection by allowing direct touch interaction with the magnified object proxies reflected on the panel view. A repeated-measures ANOVA on the results of a follow-up experiment on multitarget selection with 23 participants showed that TouchView outperformed the current market-dominating freehand raycasting technique, Hybrid Ray, in terms of task performance, perceived workload, and preference. User behavior was also analyzed to understand the underlying reasons for these improvements. The proposed technique can be used in VR UI applications to enhance the selection of distant objects, especially for cases with frequent view shifts.

Fourier‐Based Action Recognition for Wildlife Behavior Quantification with Event Cameras

Event cameras are novel bioinspired vision sensors that measure pixel‐wise brightness changes asynchronously instead of images at a given frame rate. They offer promising advantages, namely, a high dynamic range, low latency, and minimal motion blur. Modern computer vision algorithms often rely on artificial neural network approaches, which require image‐like representations of the data and cannot fully exploit the characteristics of event data. Herein, approaches to action recognition based on the Fourier transform are proposed. The approaches are intended to recognize oscillating motion patterns commonly present in nature. In particular, the approaches are applied to a recent dataset of breeding penguins annotated for “ecstatic display,” a behavior where the observed penguins flap their wings at a certain frequency. It is found that the approaches are both simple and effective, producing slightly lower results than a deep neural network (DNN) while relying just on a tiny fraction of the parameters compared to the DNN (five orders of magnitude fewer parameters). They work well despite the uncontrolled, diverse data present in the dataset. It is hoped that this work opens a new perspective on event‐based processing and action recognition.

Elevating Facial Expression Detection: Empowered by VGG-19 and Weight-Normalized Gradient Boost Algorithm

Facial expression recognition termed as FER, is one of the vital tasks which is able in mimicking the human ability and wellness. Face expression and other gestures delivered via face are some of the important aspects of conveying non-verbal communications which plays a vital role in interpersonal relations. This domain has a higher scope of research due to enhancing computer vision and human-computer interaction. With respect and consideration to these domain interest, DL has gained a reasonable approach in performing FER. Some of the state-of-art approaches encompassing the ML approaches have faced several cases of laybacks such as overfitting, computational complexity, less adaptable to higher and big datasets. Thus, considering these laybacks and in achieving accurate FER in aspects of bringing proximal levels of FER, the current approach deployed DL methods for both feature extraction from the input image and in classifying them. Deep multi-level feature extraction is performed using the VGG-19 model and Weight Normalized gradient boosting algorithm is adapted for classifying these expressions using the FER13 dataset. This dataset constitutes the input images, which are in ranges of 28,709 sample image data and about 3,589 image data for test. These input images are initially pre-processed for obtaining better accuracy rates when performing feature extraction and classifications. The complete model in effective FER is evaluated using the performance metrics comprising Accuracy (98%), Recall (99%), F1-score (98%) and the precision rates (97%). This analysis of the performance will aid in affirming the overall efficacy of the proposed system.

Moving towards the use of artificial intelligence in pain management.

While the development of artificial intelligence (AI) technologies in medicine has been significant, their application to acute and chronic pain management has not been well characterized. This systematic review aims to provide an overview of the current state of AI in acute and chronic pain management. This review was registered with PROSPERO (ID# CRD42022307017), the international registry for systematic reviews. The search strategy was prepared by a librarian and run in four electronic databases (Embase, Medline, Central, and Web of Science). Collected articles were screened by two reviewers. Included studies described the use of AI for acute and chronic pain management. From the 17,601 records identified in the initial search, 197 were included in this review. Identified applications of AI were described for treatment planning as well as treatment delivery. Described uses include prediction of pain, forecasting of individualized responses to treatment, treatment regimen tailoring, image-guidance for procedural interventions and self-management tools. Multiple domains of AI were used including machine learning, computer vision, fuzzy logic, natural language processing and expert systems. There is growing literature regarding applications of AI for pain management, and their clinical use holds potential for improving patient outcomes. However, multiple barriers to their clinical integration remain including lack validation of such applications in diverse patient populations, missing infrastructure to support these tools and limited provider understanding of AI. This review characterizes current applications of AI for pain management and discusses barriers to their clinical integration. Our findings support continuing efforts directed towards establishing comprehensive systems that integrate AI throughout the patient care continuum.

RS-DETR: An Improved Remote Sensing Object Detection Model Based on RT-DETR

Object detection is a fundamental task in computer vision. Recently, deep-learning-based object detection has made significant progress. However, due to large variations in target scale, the predominance of small targets, and complex backgrounds in remote sensing imagery, remote sensing object detection still faces challenges, including low detection accuracy, poor real-time performance, high missed detection rates, and high false detection rates in practical applications. To enhance remote sensing target detection performance, this study proposes a new model, the remote sensing detection transformer (RS-DETR). First, we incorporate cascaded group attention (CGA) into the attention-driven feature interaction module. By capturing features at different levels, it enhances the interaction between features through cascading and improves computational efficiency. Additionally, we propose an enhanced bidirectional feature pyramid network (EBiFPN) to facilitate multi-scale feature fusion. By integrating features across multiple scales, it improves object detection accuracy and robustness. Finally, we propose a novel bounding box regression loss function, Focaler-GIoU, which makes the model focus more on difficult samples, improving detection performance for small and overlapping targets. Experimental results on the satellite imagery multi-vehicles dataset (SIMD) and the high-resolution remote sensing object detection (TGRS-HRRSD) dataset show that the improved algorithm achieved mean average precision (mAP) of 78.2% and 91.6% at an intersection over union threshold of 0.5, respectively, which is an improvement of 2.0% and 1.5% over the baseline model. This result demonstrates the effectiveness and robustness of our proposed method for remote sensing image object detection.

DeepFake Image Detection: Fake Image Detection using CNNs and GANs Algorithm

Deep learning is a powerful and versatile technique that has seen extensive applications in areas such as natural language processing, machine learning, and computer vision. Among its most recent applications is the generation of deepfakes, which are high-quality, realistic altered videos or images that have garnered significant attention. While innovative uses of deepfake technology are being explored, its potential for misuse has raised serious concerns. Harmful applications, such as spreading fake news, creating celebrity pornography, financial fraud, and revenge pornography, have become increasingly prevalent in the digital age. As a result, public figures, including celebrities and politicians, face heightened risks from deepfake content. In response, substantial research has been conducted to explore the mechanics behind deepfakes, leading to the development of various deep learning-based algorithms for their detection. This study provides a comprehensive review of deepfake creation and detection techniques, focusing on different deep learning approaches. Additionally, it discusses the limitations of existing methods and the availability of datasets for research. The lack of a highly accurate and fully automated deepfake detection system presents a significant challenge as the ease of generating and distributing such content continues to grow. Nonetheless, recent efforts in deep learning have shown promising results, surpassing traditional detection methods.

An artificial intelligence-enabled consumables tracking system for medical laboratories

Abstract The medical laboratory plays a crucial role within a hospital setting and is responsible for the examination and analysis of patient specimens to accurately diagnose various ailments. The burden on medical laboratory personnel has significantly increased, particularly in the context of the ongoing global COVID-19 pandemic. Worldwide, the implementation of comprehensive and extended COVID-19 screening programs has placed a significant strain on healthcare professionals. This burden has led to exhaustion among medical employees, limiting their ability to effectively track laboratory resources, such as medical equipment and consumables. Therefore, this study proposed an artificial intelligence (AI)-based solution that contributes to a more efficient and less labor-intensive workflow for medical workers in laboratory settings. With the ultimate goal to reduce the burden on healthcare providers by streamlining the process of monitoring and managing these resources, the objective of this study is to design and develop an AI-based system for consumables tracking in medical laboratories. In this work, the effectiveness of two object detection models, namely, YOLOv5x6 and YOLOv8l, for the administration of consumables in medical laboratories was evaluated and analyzed. A total of 570 photographs were used to create the dataset, capturing the objects in a variety of settings. The findings indicate that both detection models demonstrate a notable capability to achieve a high mean average precision. This underscores the effectiveness of computer vision in the context of consumable goods detection scenarios and provides a reference for the application of real-time detection models in tracking systems within medical laboratories.

Investigating the Association of Computer Vision Syndrome with Forward-Headed and Kyphotic Posture Among Undergraduate Information Technology Student

Introduction: Computers and visual display devices have become indispensable in our daily lives. The frequent use of these devices has led to the emergence of computer vision syndrome (CVS). Coupled with poor ergonomics, musculoskeletal diseases are on the rise due to excessive computer use. Accordingly, this study investigates the association between CVS and forward-headed and kyphotic posture in undergraduate IT students. Materials and Methods: We conducted an observational cross-sectional study involving 300 undergraduate IT students. We used a modified CVS questionnaire to assess CVS and the Kinovea software to diagnose forward head posture. We also used the occiput wall distance test to assess kyphosis. Results: In this study, a majority of undergraduate students (88%, 264 out of 300) exhibited CVS, resulting in forward head posture and kyphosis. There was a significant association between CVS, forward head posture, and kyphosis. Conclusion: Computer vision syndrome is strongly associated with forward head posture and kyphosis. Prolonged computer device usage leads to changes in body posture, affecting the neck, shoulders, elbows, wrists/hands, upper back, and lower back.

Image modeling algorithm for environment design based on augmented and virtual reality technologies

Abstract With the continuous development of computer vision technology, the image modeling of environmental design has also received more and more attention. This study proposed an image modeling algorithm for environment design based on augmented reality (AR) and virtual reality (VR) technologies. On the system architecture, the environment design image modeling algorithm based on AR and VR technologies has provided users with an immersive 3D simulation space environment. It also has provided users with a good interactive interface through AR technology, enabling users to interact with computers in a realistic virtual three-dimensional environment and interactive operation. This study verified the algorithm from three aspects: time, authenticity, and interactivity. When facing object A, the modeling time values of the traditional environmental design image modeling method and the environmental design image modeling method proposed in this study were 12.14 and 8.72 s, respectively. The modeling authenticity reached 87.51 and 98.57%, respectively, and the interactivity reached 79.14 and 96.45%, respectively. Some columns of data have proved that the environmental design image modeling method proposed in this study was superior to the traditional environmental design image modeling method. In addition, this study also verified the system model from three aspects: stability, running smoothness, and performance loss. When the number of image modeling in the environment design was 2, the stability, running smoothness, and performance loss of the system were 99.37, 99.25 and 0.75%, respectively. The experimental data again proved that the algorithm system model proposed in this study was feasible and worthy of further application.