检索结果-内蒙古大学图书馆

Fλ/d as a predictor of image classification algorithm performance

APPLIED OPTICS 2025年第4期64卷 845-854页

作者： Hixson, Jonathan g. Teaney, Brian Finch, Michael f. Nehmetallah, George Driggers, Ronald US Army Combat Capabil Dev Command DEVCOM C5ISR Ctr RTI Aberdeen Proving Ground NY 21005 USA Catholic Univ Amer EECS Dept 620 Michigan Ave NE Washington DC 20064 USA Univ Arizona Wyant Coll Opt Sci Infrared Syst Grp Tucson AZ 85721 USA

This research studied the effect of variations in a sensor's F lambda/d metric value (FLD) on the performance of machine learning algorithms such as the YOLO (You Only Look Once) algorithm for object classification. The YOLO_v3 and YOLO_v10 algorithms were trained using static imagery provided in the commonly available training dataset provided by Teledyne FLIR systems. image processing techniques were used to degrade image quality of the test dataset also provided by Teledyne FLIR systems, simulating detector-limited to optics-limited performance, which results in a variation of the FLD metric between 0.339 and 7.98. The degraded test set was used to evaluate the performance of YOLO_v3 and YOLO_v10 for object classification and relate the FLD metric to the probability of detection. Results of YOLO_v3 and YOLO_v10 are presented for the varying levels of image degradation. A summary of the results is discussed along with recommendations for evaluating an algorithm's performance using a sensor's FLD metric value. (c) 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

关键词： image quality image sensors Imaging systems Imaging techniques Machine learning Sensor performance

来源：评论

学校读者我要写书评

暂无评论

Optimized Residual Attention Based Generalized Adversarial Network for COvID-19 Classification Using Chest CT images

引用

COMPUTATIONAL INTELLIGENCE 2025年第2期41卷

作者： Sarvari, A. v. P. Sridevi, K. GITAM Deemed Univ Dept Elect Elect & Commun Engn Visakhapatnam Andhra Prades India

The early detection and classification of COvID-19 is crucial for disease diagnosis and control. To reduce the need for medical professionals, fast and accurate detection approaches for COvID-19 are required. Due to environmental concerns, the quality of the image gets degraded. Compared with reverse-transcription polymerase chain reaction (RT-PCR), chest computed tomography (CT) imaging may be a significantly more trustworthy, useful, and rapid technique to classify and evaluate COvID-19. Thus, the performance of the deep learning (DL) techniques is diminished. Therefore, a CT image-based hybrid DL technology is presented in this article for the classification of COvID-19 disease as COvID or non-COvID or pneumonia. Initially, in the pre-processing stage, the hybrid nonlocal moment bilateral filtering (Hybrid NMBF) technique is introduced for image de-noising and re-sizing. After pre-processing, the image is fed into the feature extraction phase. Gray-level covariance matrices (GLCM) technique is used to extract the relevant features and reduce feature dimensionality issues. For the feature selection process, the enhanced Archimedes optimization algorithm (EAOA) is introduced to select optimal features. The residual channel attention-generative adversarial network (RCA-GAN) technique is introduced for image classification. Here, the hyperparameter of the network is tuned using the Sandpiper optimization (SPO) algorithm to optimize the loss function. The data set used in this research is COvID-CT-machine learning deep learning (MD), and the performance is analyzed using the MATLAB tool. In the experimental scenario, the proposed system achieves 98.3% accuracy, 98.7% specificity, 99.4% sensitivity, 97.4% F-score, and 96.1% kappa. The attained results prove that the proposed system works better than the traditional techniques.

关键词： chest CT image COvID-19 deep learning nonlocal moment bilateral filtering residual channel attention-general adversarial network

来源：评论

学校读者我要写书评

暂无评论

TranGDeepSC: Leveraging viT knowledge in CNN-based semantic communication system

引用

ICT EXPRESS 2025年第2期11卷 335-340页

作者： Do, Tung Son Truong, Thanh Phung Do, Quang Tuan Cho, Sungrae Chung Ang Univ Sch Comp Sci & Engn Seoul 06974 South Korea

This paper introduces TranGDeepSC, a lightweight CNN-based deep semantic communication (DeepSC) system that leverages vision Transformer (viT) knowledge through co-training to enhance image transmission. Evaluated on CIFAR-100 across various SNRs, TranGDeepSC demonstrates competitive performance with viTDeepSC, and outperforms SemviT and ADJSCC-v in image quality, particularly in low-SNR environments. Notably, it offers substantial gains in efficiency: 92.8% fewer parameters than ADJSCC-v, 72.0% lower energy use, and 48% faster processing than viTDeepSC. These advantages make TranGDeepSC well-suited for resource-constrained applications in next-generation communication systems, including 6G, IoT, and real-time multimedia streaming.

关键词： CNN Lightweight Semantic communication 6G

来源：评论

学校读者我要写书评

暂无评论

White blood cell classification using multi-hop attention graph neural networks

引用

EXPERT systemS WITH APPLICATIONS 2025年 272卷

作者： Duc, Minh Ly Bilik, Petr Martinek, Radek Van Lang Univ Fac Commerce Ho Chi Minh City Vietnam VSB Tech Univ Ostrava Fac Elect Engn & Comp Sci Dept Cybernet & Biomed Engn 17 Listopadu 15 Ostrava 70800 Czech Republic

The process of creating blood cells (hematopoiesis) occurs in the bone marrow, where Hematopoietic Stem Cells (HSCs) are located. The division and differentiation of hematopoietic stem cells are tightly regulated to ensure a balance between blood cell lineages. Disturbances in the process can lead to blood diseases such as anemia, high White Blood Cell (WBC) count, or thrombocytopenia. Detecting malignant leukemia cells based on images is crucial in diagnosing and treating leukemia, helping doctors make accurate diagnoses, and providing appropriate treatment. The author proposes a new method for recognizing and classifying WBC images using the Multi-hop Attention Graph Neural Networks method. The YOLO-v10 method is used for object detection and image preprocessing through the Centre Net network architecture. The Salp Swarm Optimization (SSO) method is deployed to select the features of the WBC images optimally and put the image features into each node in the architecture of the Graph Neural Network (GNN) model to perform classification. The dataset used has an image quality of approximately 42 pixels per 1 mu m resolution with a total of 16,027 annotated White Blood Cell images classified into 9 types of WBC with characteristic images of clinically significant pathologies. The classification accuracy of the system of the YLSSOGNN model is 99.18%, and the classification accuracy of the system of the YLGNN model is 99.03 %. The WBC image recognition and classification model using the post-learning method has a GNN architecture with object recognition function using the YOLO-v10 method and feature extraction and optimization using the SSO method and performs WBC image classification using Multi-hop Attention Graph Neural Networks model, which helps to bring high performance and can apply the model to other types of image objects.

关键词： White blood cell Graph neural network YOLO-v10 Salp Swarm Optimization

来源：评论

学校读者我要写书评

暂无评论

Deep learning-driven UAv vision for automated road crack detection and classification

引用

NONDESTRUCTIvE TESTING AND EvALUATION 2025年

作者： Rathod, vaishnavee v. Rana, Dipti P. Mehta, Rupa G. Sardar Vallabhbhai Natl Inst Technol Dept Comp Sci Engn Surat Gujarat India

Unmanned aerial vehicles are increasingly utilised for monitoring and inspecting critical infrastructure such as power generation grids, oil and gas pipelines and roads. One key task in road maintenance is the detection of cracks, which is crucial for ensuring road safety. Manual detection of cracks is time-consuming and prone to errors, highlighting the need for automated solutions. This study presents an automated method for road crack detection and classification using a hybrid deep learning technique. The proposed approach integrates a pyramid vision transformer and ConvMixer models for feature extraction, enabling the system to learn complex patterns in crack images. image pre-processing is first performed using a median filtering technique to enhance image quality. The detection and classification of cracks are then carried out using an Elman neural network model, with its hyperparameters optimised through an improved black widow optimisation algorithm. Extensive simulations demonstrate that the proposed method outperforms other deep learning models in terms of performance, providing a reliable and efficient solution for automated crack detection.

关键词： Unmanned aerial vehicles road crack detection black widow optimization feature fusion deep learning

来源：评论

学校读者我要写书评

暂无评论

Research on the Application of Multi-Channel Display Technology in virtual Scene image Generation 5

Research on the Application of Multi-Channel Display Technol...

引用

5th IEEE International conference on Power, Electronics and Computer Applications, ICPECA 2025

作者： Liu, Yuanlin Wang, Lingling Haikou University of Economics Haikou571127 China

ISBN: (纸本)9798331533694

This study proposes an innovative algorithm based on DCNN and multi-channel image fusion, aiming to improve the quality and efficiency of virtual scene image generation. The algorithm extracts depth information and texture features in virtual scenes through DCNN, and combines multi-channel image fusion technology to optimize and fuse the information of different image channels (such as color, depth, normal, etc.), thereby significantly improving the image rendering effect. Through adaptive weight adjustment and image fusion strategy, the image detail performance is optimized, especially in scenes with complex lighting and rich textures, the rendering quality is significantly improved. Experimental results show that the image generation system based on this algorithm improves image quality by 15%, rendering speed by 25%, and computing resource consumption by 20%. When processing highly complex virtual scenes, the system can achieve real-time rendering and meet the performance requirements of vR and AR applications. This study provides a new and efficient solution for image generation in virtual reality and augmented reality, and demonstrates the great potential of combining multi-channel display technology with deep learning in improving rendering effects. © 2025 IEEE.

关键词： virtual environments

来源：评论

学校读者我要写书评

暂无评论

variational Onsager Neural Network optimized with Golden search optimization algorithm fostered for lung disease detection system in IoT

引用

BIOMEDICAL SIGNAL PROCESSING AND CONTROL 2025年 108卷

作者： Sudha, G. Angayarkanni, v. Kanagavalli, K. R. Pattewar, Tareek Muthayammal Engn Coll Dept Biomed Engn Rasipuram 637408 Tamil Nadu India SRM Inst Sci & Technol Fac Engn &Technol Dept Comp Technol Kattankulathur 603203 Tamil Nadu India Govt Arts & Sci Coll Dept Comp Sci Sivakasi 626124 Tamil Nadu India Vishwakarma Univ Dept Comp Sci & Engn Pune India

Pneumonia causes a high rate of newborn morbidity and mortality. The challenge is accurately identifies respiratory disorders while overcoming the limitations of existing technologies such as low accuracy, delayed response, and restricted scalability. To overcome this complication, variational Onsager Neural Network optimized with Golden search optimization algorithm fostered for Lung Disease Detection system in IoT (LDD-vONNCXR-IoT) is proposed. Initially, input CXR images are gathered from chest-X-ray Dataset. Then, pre-process the input CXR images using Two-way Recursive filtering (TWRF) for normalizing image and increasing the quality of the images. Afterwards, the preprocessed image is supplied to the feature extraction. Adaptive Synchro Extracting Transform (ASET) is employed to extract the statistical features. Finally, the extracted features are fed into variational Onsager Neural Networks (vONN) which classifies the input CXR image into normal and pneumonia. The Golden Search Optimization Algorithm (GSOA) is used to optimize vONN that accurately detects the Lung Disease. The proposed LDD-vONN-CXR-IoT method is implemented. The performance metrics, like precision, accuracy, F1-score, Sensitivity, specificity, Error rate, ROC, computational time are examined. The proposed LDD-vONN-CXR-IoT approach attains 99.57%, 98.46%, and 98.13% for accuracy, F1 score, and precision respectively. These outcomes prove that this method for the Lung Disease Detection system in IoT is effectual tool to assist in clinical diagnosis. This method allows expertise to acquire exact results, thus providing the proper treatment.

关键词： Adaptive Synchrony extracting Transform Chest-X-ray Dataset Golden Search Optimization Algorithm Two-way Recursive filtering variational Onsager Neural Network Chest Radiograph image

来源：评论

学校读者我要写书评

暂无评论

Comparative analysis of optical system configurations for enhanced retinal imaging in smartphone-based fundus cameras

Comparative analysis of optical system configurations for en...

引用

2024 International conference on Interdisciplinary Physics, ICIPs 2024

作者： Rosyadah, Aprina Nasution, Aulia M.T. Surabaya Indonesia

This paper presents a comparative study on the optical system design for a smartphone-based fundus camera to enhance retinal imaging quality. Three optical system configurations were evaluated using ANSYS Zemax OpticStudio 2024 R1 software: Design I (20D aspheric lens), Design II (40D aspheric lens), and Design III (60D aspheric lens). The optical systems were evaluated by analysing: (i) the illumination path from the light source to the retina, and (ii) the imaging path from the retina to the smartphone camera's sensor. The results were evaluated using image metric quantities, including the spot diagram, transverse ray fan plot, and TTF MTF vs Field, to represent the image quality of each optical system. Additionally, the illumination and imaging paths were analysed to assess the overall image quality. The findings indicated that Design III, with the 60D aspheric lens, demonstrated superior image quality in terms of spot diagram, ray fan plot, and MTF performance, making it the most effective design. © 2025 Institute of Physics Publishing. All rights reserved.

关键词： Ophthalmology

来源：评论

学校读者我要写书评

暂无评论

A novel opto-tactile sensing approach to enhance the handling of soft fruit

引用

COMPUTERS AND ELECTRONICS IN AGRICULTURE 2025年 235卷

作者： Ameur, Mohamed Adlan Ait El-Sayed, Amr M. Yan, Xiu T. Mehnen, Jorn Maier, Anja M. Univ Strathclyde Design Mfg & Engn Management Glasgow Scotland Assiut Univ Fac Engn Mechatron Engn Dept Assiut 71516 Egypt Glasgow Caledonian Univ Dept Mech Engn Cowcaddens Rd Glasgow G4 0BA Scotland

In agricultural settings, handling of soft fruit is critical to ensuring quality and safety. This study introduces a novel opto-tactile sensing approach designed to enhance the handling and assessment of soft fruit, with a case example of strawberries. Our approach utilises a Robotiq 2F-85 gripper equipped with the DIGIT vision-Based Tactile Sensor (vBTS) and attached to a Universal Robot UR10e. In contrast to force-based approaches, we introduce a novel purely image-based processing software pipeline for quantifying localised surface deformations in soft fruit. The system integrates fast and explainable image processing techniques applying image differencing, denoising, K-means clustering for unsupervised classification, morphological operations, and connected components analysis (CCA) to quantify surface deformations accurately. A calibration of the image processing pipeline using a rubber ball showed that the system effectively captured and analysed the rubber ball's surface deformations, benefiting from its uniform elasticity and predictable response to compression. As a soft fruit case example, the image processing pipeline was subsequently applied to strawberries, blueberries, and raspberries, demonstrating that the calibration parameters derived from the rubber ball could effectively assess surface deformations in soft fruits. Despite the complex, nonlinear deformation characteristics inherent to strawberries, blueberries, and raspberries, the pipeline exhibited robust performance, capturing and quantifying subtle surface changes. These findings underscore the system's capacity for precise deformation analysis in delicate materials, offering major potential for further refinement and adaptation. This novel approach of proposing and testing an image processing pipeline lays the groundwork for enhancing the handling and assessment of materials with intricate mechanical properties, paving the way for broader applications in sensitive agricultural and industrial

关键词： Robotic gripper vision-Based Tactile Sensor (vBTS) image Processing Machine Learning Soft Fruit Handling Robotic manipulation

来源：评论

学校读者我要写书评

暂无评论

Diffusion Model Based image Reconstruction in Lensless Imaging

Diffusion Model Based Image Reconstruction in Lensless Imagi...

引用

2025 IEEE International conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： verma, Ashish Boominathan, vivek veeraraghavan, Ashok Seelamantula, Chandra Sekhar Department of Electrical Engineering Indian Institute of Science Bengaluru560 012 India Department of Electrical and Computer Engineering Rice University HoustonTX77005 United States

ISBN: (纸本)9798350368741

Lensless imaging systems eliminate the need for lenses by employing an encoding element to multiplex incident light signals, which are then captured directly onto a bare camera sensor. They present a promising alternative to traditional lens-based imaging systems by offering significant advantages in terms of compactness, versatility, and cost. Due to the multiplexed nature of measurements, image reconstruction takes place computationally. However, existing techniques for image reconstruction in lensless imaging fall short of the image quality offered by traditional lens-based imaging. In this work, we consider the application of diffusion models, a class of deep generative models, for image reconstruction in a lensless imaging modality. These models currently achieve state-of-the-art performance in image generation. Specifically, we focus on the PhlatCam lensless system, which consists of a coded phase mask as the encoding element placed close to the camera sensor. We use a ControlNet based diffusion model to improve the perceptual quality of image reconstruction. The performance is measured in terms of peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and learned perceptual image patch similarity (LPIPS). The proposed method improves the performance in these metrics for synthetic measurements. For real measurements, the improvement in image quality comes at the expense of a small bias in color, which is attributed to the generative nature of the diffusion prior itself. © 2025 IEEE.

关键词： Computational imaging diffusion model generative model lensless imaging stable diffusion

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：