检索结果-内蒙古大学图书馆

Medical Imaging Conference - image processing

作者： Sengupta, Sourya Athwale, Akshay Gulati, Tanmay Zelek, John Lakshminarayanan, Vasudevan Univ Waterloo Sch Optometry & Vis Sci Theoret & Expt Epistemol Lab Waterloo ON Canada Univ Waterloo Dept Syst Design Engn Waterloo ON Canada IIT Dhanbad Dept Appl Math Dhanbad Bihar India Manipal Insitute Technol Dept Comp Sci Manipal India

ISBN: (纸本)9781510633940

Medical imaging datasets typically do not contain many training images and are usually not sufficient for training deep learning networks. We propose a deep residual variational auto-encoder and a generative adversarial network based approach that can generate a synthetic retinal fundus image dataset with corresponding blood vessel annotations. In terms of structural statistics comparison of real and artificial our model performed better than existing methods. The generated blood vessel structures achieved a structural similarity value of 0.74 and the artificial dataset achieved a sensitivity of 0.84 and specificity of 0.97 for the blood vessel segmentation task. The successful application of generative models for the generation of synthetic medical data will not only help to mitigate the small dataset problem but will also address the privacy concerns associated with such medical datasets.

关键词： Fundus RSVAE GAN image Synthesis Retina Ophthalmology

来源：评论

学校读者我要写书评

暂无评论

Predicting Parkinson’s Disease: An Explainable AI approach with Feature Analysis

引用

Procedia Computer Science 2025年 259卷 1773-1782页

作者： Madhurjya Rabha Krishnav Deka deepratim Saikia Junali Jasmine Jena Sudhansu Shekhar Patra Mahendra Kumar Gourisaria School of Computer Engineering KIIT Deemed to be University Bhubaneswar-751024 Odisha India School of Computer Applications KIIT Deemed to be University Bhubaneswar-751024 Odisha India

Parkinson’s disease (PD) is a degenerative neurological condition that significantly lowers a person’s standard of living. Symptom management and better patient outcomes depend on early diagnosis and intervention. To forecast Parkinson’s disease, this study looks into the use of various ML models, such as Decision Trees, XGBoost, AdaBoost etc. To improve model interpretability, the study highlights the significance of Explainable Artificial Intelligence (XAI) approaches. The essential elements that had a substantial impact on the models’ predictions were found by using feature importance analysis. In addition to illuminating the disease’s fundamental causes, this study offers medical professionals’ practical insights to direct clinical judgment. The integration of machine learning with XAI techniques addresses the critical need for transparency in AI applications within healthcare, ensuring that practitioners can trust and understand the tools they use. Among the classifiers evaluated, the AdaBoost classifier demonstrated exceptional performance, achieving the highest accuracy of 98.34%. This result highlights the strong performance of the AdaBoost algorithm in accurately predicting Parkinson’s disease, demonstrating its suitability and effectiveness for this application. These algorithms are computationally less expensive than the deep learning algorithms and hence could be efficiently used for real-time systems. Also in this paper the contribution of features have been mapped to the accuracy, which shows the explainability of features. The work aims to set the stage for future developments in neurodegenerative disease predictive modeling through thorough analysis and validation.

关键词： ML Models Parkinson’s Disease Feature Extraction Explainable AI image processing

来源：评论

学校读者我要写书评

暂无评论

real time social distance detection using deep learning 1

Real time social distance detection using Deep Learning

引用

1st International Conference on Computational Intelligence and Sustainable Engineering Solution, CISES 2022

作者： Salagrama, Shailaja Kumar, H. Hemant Nikitha, R. Prasanna, G. Sharma, Kanhaiya Awasthi, Shashank University of the Cumberland's Information Technology WilliamsburgKY United States Anurag University Computer Science and Engineering Hyderabad India Anurag University Computer Science and Text Engineering Hyderabad India G.L. Bajaj I.T.M. Computer Science and Engineering Greater Noida India

ISBN: (数字)9781665480048

ISBN: (纸本)9781665480048

In the Covid-19 pandemic, if residents do not take action to prevent the virus from spreading, the process of softening the curve of the coronavirus will be complicated in the face of the worldwide Covid-19 scenario. Without vaccination, the only method to combat the disease is social isolation. The proposed system employs the You Only Look Once, Version 3 (YOLOv3) object detection model to identify persons in the background and bind boxes around them, and assign IDs for in-depth tracking of recognized people. This study focuses on public space surveillance and determining whether or not people maintain social distance as per Covid-19 guidelines. YOLOv3 is an efficient tracking method that produces positive results with a moderate mean Average Precision(MAP) and Frame Per Second (FPS) score for monitoring community deviations in real-time. In this study, YOLOv3 is used for object capture, and the OpenCV library is used for image processing. Proposed work is helpful in areas where big crowds are expected, such as retail malls, movie theatres, railway stations, airports, and public places. © 2022 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

E2BNet: MAC-free yet accurate 2-level binarized neural network accelerator for embedded systems

引用

JOURNAL OF real-time image processing 2021年第4期18卷 1285-1299页

作者： Mirsalari, Seyed Ahmad Nazari, Najmeh Ansarmohammadi, Seyed Ali Salehi, Mostafa E. Ghiasi, Soheil Univ Tehran Sch Elect & Comp Engn Tehran Iran Inst Res Fundamental Sci IPM Sch Comp Sci Tehran Iran Univ Calif Davis Elect & Comp Engn Dept Davis CA 95616 USA

deep neural networks are widely used in computer vision, pattern recognition, and speech recognition and achieve high accuracy at the cost of remarkable computation. High computational complexity and memory accesses of such networks create a big challenge for using them in resource-limited and low-power embedded systems. Several binary neural networks have been proposed that exploit only 1-bit values for both weights and activations. Binary neural networks substitute complex multiply-accumulation operations with bitwise logic operations to reduce computations and memory usage. However, these quantized neural networks suffer from accuracy loss, especially in big datasets. In this paper, we introduce a quantized neural network with 2-bit weights and activations that is more accurate compared to the state-of-the-art quantized neural networks, and also the accuracy is close to the full precision neural networks. Moreover, we propose E2BNet, an efficient MAC-free hardware architecture that increases power efficiency and throughput/W about 3.6 x and 1.5 x , respectively, compared to the state-of-the-art quantized neural networks. E2BNet processes more than 500 images/s on the imageNet dataset that not only meet real-time requirements of images/video processing but also can be deployed on high frame rate video applications.

关键词： Binary neural network deep learning Embedded systems real-time Power efficiency

来源：评论

学校读者我要写书评

暂无评论

A feature extraction based support vector machine model for rectal cancer T-stage prediction using MRI images

引用

MULtimeDIA TOOLS AND APPLICATIONS 2021年第20期80卷 30907-30917页

作者： Wang, Yizhang Gong, Tingting Hassan, Muhammad Li, Qiang Huang, Sa Zhou, You Jilin Univ Coll Comp Sci & Technol Changchun Peoples R China Yangzhou Univ Coll Informat Engn Yangzhou Jiangsu Peoples R China Second Hosp Jilin Univ Dept Radiol Changchun Peoples R China Knowledge Engn Minist Educ Key Lab Symbol Computat Changchun Peoples R China Second Hosp Jilin Univ Changchun Peoples R China

Accurate clinical cancer T-stage diagnosis is crucial for effective treatment. However, it is difficult, time-consuming, and laborious for physicians to recognize T-stage manually using rectal Magnetic Resonance Imaging (MRI) images. Machine learning methods have played important roles in medical image processing. With the goal of automatic rectal cancer T-stage prediction, we train the proposed Feature Extraction based Support Vector Machine (FE-SVM) model with the newly acquired dataset consisting of 147 patients' MRI images with primary rectal cancer. Our method adapts SVM as the training framework as SVM is effective enough for dealing with small datasets as opposed to state-of-the-art deep learning models. FE-SVM firstly extracts image similarity as an initial feature because the feature of image similarity can better reflect the differences among various types of MRI images, and the final 10-dimensional features are obtained by a 5-layers Autoencoder. To evaluate the performance of FE-SVM, we compared it with six benchmark models: CNN, Alexnet, Resnet18, Resnet50, Capsule Network, and Random Forest. FE-SVM outperforms these state-of-the-art models with significant evaluation scores.

关键词： Feature extraction MRI image Machine learning

来源：评论

学校读者我要写书评

暂无评论

Semantic segmentation of superficial layer in intracoronary optical coherence tomography based on cropping-merging and deep learning

引用

INFRARED PHYSICS & TECHNOLOGY 2023年第1期129卷

作者： Ren, Xinbo Zhao, Yuxia Fan, Jingbo Wu, Haiyuan Chen, Qian Kubo, Takashi Shangluo Univ Dept Informat Management Shangluo 726000 Peoples R China Shangluo Univ Sch Math & Comp Sci Shangluo 726000 Peoples R China Wakayama Univ Grad Sch Syst Engn Wakayama Japan SenseTime Japan Ltd Sch Math & Comp Sci Kyoto Japan Wakayama Med Univ Dept Cardiovasc Med Wakayama Japan

Manual labelling and qualitative analysis of lesion tissues in vessel optical coherence tomography (OCT) images are time-consuming and laborious for cardiovascular specialists. To beat these issues, a semantic segmentation method was presented based on deep learning on the OCT image analysis of the human vessel to reduce the diagnosis pressure of cardiovascular doctors. An outer border of ROI was segmented based on the level-set method to obtain the visible superficial layer containing useful information. And then, cropping square patches from the ROI with its patch center pixel inside the ROI and using patches as the input data. To reuse the preceding layers' feature maps, we used the dense block to replace the normal convolution layer, and simul-taneously, employed a skip-connection from m the down-sampling path to the up-sampling path to keep the spatial information. With the advantage of SegNet on semantic segmentation, a dense-block-SegNet (DBSegNet) is constructed to complete the pixel-level segmentation. Training and testing were executed on 7 datasets (22, 210 images) to assess the model. Tenfold cross-validation method was implemented to measure the classification outcomes and the semantic segmentation capacity of our model. In deep learning experiment, sensitivities for calcified, fibrous and lipid plaques were 91.81 +/- 3.60 %, 92.81 +/- 2.72 % and 91.78 +/- 1.62 %, respectively. A 3-D volume was created to insert each prediction slice along the depth axis to compute the maximum type number of each pixel in order to identify the final type of each pixel. Post-processing was used to refine the classification findings in order to eliminate classification errors. Semantic segmentation neural network based on the cropped input data and feature reusing was a feasible approach for the vessel tissue pixel-classification of IVOCT image. The proposed method has the potential to become a useful tool for specialists in analysing lesion kinds and determining clinical

关键词： IVOCT Semantic segmentation Superficial layer Lesion plaques deep learning

来源：评论

学校读者我要写书评

暂无评论

DCAVN: Cervical cancer prediction and classification using deep convolutional and variational autoencoder network

引用

MULtimeDIA TOOLS AND APPLICATIONS 2021年第20期80卷 30399-30415页

作者： Khamparia, Aditya Gupta, deepak Rodrigues, Joel J. P. C. de Albuquerque, Victor Hugo C. Lovely Profess Univ Sch Comp Sci & Engn Phagwara Punjab India Maharaja Agrasen Inst Technol New Delhi India Univ Fed Piaui Teresina PI Brazil Inst Telecommun P-1049001 Lisbon Portugal Univ Fortaleza Grad Program Appl Informat Fortaleza Ceara Brazil

Early detection, early diagnosis and classification of the cancer type facilitates faster disease management of patients. Cervical cancer is fourth most pervasive cancer type which affects life of many people worldwide. The intent of this study is to automate cancer diagnosis and classification through deep learning techniques to ensure patients health condition progress timely. For this research, Herlev dataset was utilized which contains 917 benchmarked pap smear cells of cervical with 26 attributes and two target variables for training and testing phase. We have adopted combination of convolutional network with variational autoencoder for data classification. The usage of variational autoencoder reduces the dimensionality of data for further processing with involvement of softmax layer for training. The results have been obtained over 917 cancerous image type pap smear cells, where 70% (642) allocated for training and remaining 30% (275) considered for test data set. The proposed architecture achieved variational accuracy of 99.2% with 2*2 filter size and 99.4% with 3*3 filter size using different epochs. The proposed hybrid variational convolutional autoencoder approach applied first time for cervical cancer diagnosis and performed better than traditional machine learning methods.

关键词： Variational Convolution Cervical deep learning Autoencoder

来源：评论

学校读者我要写书评

暂无评论

Pedestrian Trajectory Prediction Focused on Socially Interactive Scenarios

Pedestrian Trajectory Prediction Focused on Socially Interac...

引用

International Conference on Signal processing and Communications, SPCOM

作者： Harshdeep Singh Pruthi Ranveer Singh Rahul Kala G.C Nandi Center of Intelligent Robotics Indian Institute of Information Technology Allahabad Prayagraj India Center of Autonomous Systems ABV-Indian Institute of Information Technology and Management Gwalior Gwalior India Head Center of Intelligent Robotics Indian Institute of Information Technology Allahabad Prayagraj India

ISBN: (数字)9798350350456

ISBN: (纸本)9798350350463

The real-world application of Computational Imaging and image processing in the fields like autonomous Vehicles and robot-based warehouses is immense. The guiding of these machines safely and securely is an important task. And hence the knowledge and a precise judgement on human trajectories based on the real time data collected becomes an equally important task, given humans move through complicated, crowded situations by abiding social norms. deep learning techniques recently fared better than their manually produced equivalents because they learn about human-human interactions in a more general, data-driven way. We base our work on such standard architecture and implement architectural changes. We design a new way of input encoding, which captures the long-term intent better than only taking positions or velocities. We introduce a novel Channel-wise attention mechanism using CNN in the bottleneck of LSTM encoder-decoder network. We came up with a curvature loss and also a new metric with a specially crafted dataset to implement this metric that evaluates the models on how well they could predict a turn in the right direction of the right magnitude. Experimental results show that the proposed approach outperforms popular approaches and several baselines.

关键词： Measurement Pedestrians Predictive models Signal processing real-time systems Trajectory Task analysis

来源：评论

学校读者我要写书评

暂无评论

deep-learning Based Quantification of Bovine Oocyte Quality From Microscopy images*

Deep-Learning Based Quantification of Bovine Oocyte Quality ...

引用

Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

作者： Matilde Costa Anoek Strumane Annelies Raes Ann Van Soom Danilo Babin Jan Aelterman Department of Chemical Engineering Faculty of Engineering University of Porto Porto Portugal Department of Telecommunications and Information Processing - Image Processing and Interpretation Ghent University-Imec Ghent Belgium Department of Internal Medicine Reproduction and Population Medicine Faculty of Veterinary Medicine Ghent University Merelbeke Belgium

The success rate of bovine in vitro embryo reproduction is low and highly dependent on the oocyte quality. The selection of the oocyte to be fertilized is done by the embryologists’ visual examination of oocytes. It is time-consuming, subjective, and inconsistent between specialists in the area. In this paper, a semi-automatic solution is proposed to score the quality of an immature oocyte. It consists of a deep learning model to classify oocyte competence. The model was trained and tested with real data, composed of images of immature oocytes and their label of whether they developed into blastocysts after fertilization. To the best of our knowledge, automated bovine oocyte classification was not attempted before, but experimental results show that our proposed solution is more robust and objective than specialists’ visual assessment and comparable with other works on human *** relevance— This establishes a semi-automatic real-time method to score bovine immature oocytes, based on stereo-microscopy images. Our method will significantly reduce the time of in vitro embryo production and its success.

关键词：

来源：评论

学校读者我要写书评

暂无评论

基于偏振编码图像的低空伪装目标实时检测

引用

兵工学报 2024年第5期45卷 1374-1383页

作者：沈英刘贤财王舒黄峰福州大学机械工程及自动化学院福建福州350108

偏振可以提高无人机的自主侦察能力,但易受到探测角度和目标材质的影响,从而降低偏振检测的鲁棒性。为此,提出一种基于偏振图像的低空伪装目标实时检测算法YOLO-P,采用融合多偏振方向信息的编码图像作为输入,应用三维卷积模块提取不同... 详细信息

偏振可以提高无人机的自主侦察能力,但易受到探测角度和目标材质的影响,从而降低偏振检测的鲁棒性。为此,提出一种基于偏振图像的低空伪装目标实时检测算法YOLO-P,采用融合多偏振方向信息的编码图像作为输入,应用三维卷积模块提取不同偏振方向图像之间的联系特征;引入特征增强模块对多层次特征进行进一步增强;采用跨层级特征聚合网络,充分利用不同尺度的特征信息,完成特征的有效聚合,最终联合多通道特征信息输出检测结果。构建包含10类目标的低空伪装目标偏振图像数据集PICO(Polarization image of Camouflaged Objects)。在PICO数据集上的实验结果表明,新方法可以有效检测伪装目标,mAP_(0.5:0.95)达到52.0%,mAP_(0.5)达到91.5%,检测速率达到55.0帧/s,满足实时性要求。

关键词：无人机伪装目标检测深度学习偏振成像特征增强特征聚合

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：