检索结果-内蒙古大学图书馆

Pediatric Wrist Fracture Detection Using Feature Context Excitation Modules in X-ray Images

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ju, Rui-Yang Chien, Chun-Tse Xieerke, Enkaer Chiang, Jen-Shiun Graduate Institute of Networking and Multimedia National Taiwan University Taipei City106319 Taiwan Department of Electrical and Computer Engineering Tamkang University New Taipei City251301 Taiwan College of Energy and Mechanical Engineering Shanghai University of Electric Power Shanghai201306 China

Children often suffer wrist trauma in daily life, while they usually need radiologists to analyze and interpret X-ray images before surgical treatment by surgeons. The development of deep learning has enabled neural networks to serve as computer-assisted diagnosis (CAD) tools to help doctors and experts in medical image diagnostics. Since YOLOv8 model has obtained the satisfactory success in object detection tasks, it has been applied to various fracture detection. This work introduces four variants of Feature Contexts Excitation-YOLOv8 (FCE-YOLOv8) model, each incorporating a different FCE module (i.e., modules of Squeeze-and-Excitation (SE), Global Context (GC), Gather-Excite (GE), and Gaussian Context Transformer (GCT)) to enhance the model performance. Experimental results on GRAZPEDWRI-DX dataset demonstrate that our proposed YOLOv8+GC-M3 model improves the mAP@50 value from 65.78% to 66.32%, outperforming the state-of-the-art (SOTA) model while reducing inference time. Furthermore, our proposed YOLOv8+SE-M3 model achieves the highest mAP@50 value of 67.07%, exceeding the SOTA performance. The implementation of this work is available at https://***/RuiyangJu/FCE-YOLOv8. Copyright © 2024, The Authors. All rights reserved.

关键词： Medical image processing

YOLOv9 for Fracture Detection in Pediatric Wrist Trauma X-ray Images

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chien, Chun-Tse Ju, Rui-Yang Chou, Kuang-Yi Chiang, Jen-Shiun Department of Electrical and Computer Engineering Tamkang University New Taipei City251301 Taiwan Graduate Institute of Networking and Multimedia National Taiwan University Taipei City106335 Taiwan School of Nursing National Taipei University of Nursing and Health Sciences Taipei City112303 Taiwan

The introduction of YOLOv9, the latest version of the You Only Look Once (YOLO) series, has led to its widespread adoption across various scenarios. This paper is the first to apply the YOLOv9 algorithm model to the fracture detection task as computer-assisted diagnosis (CAD) to help radiologists and surgeons to interpret X-ray images. Specifically, this paper trained the model on the GRAZPEDWRI-DX dataset and extended the training set using data augmentation techniques to improve the model performance. Experimental results demonstrate that compared to the mAP 50-95 of the current state-of-the-art (SOTA) model, the YOLOv9 model increased the value from 42.16% to 43.73%, with an improvement of 3.7%. The implementation code is publicly available at https://***/RuiyangJu/YOLOv9-Fracture-Detection. Copyright © 2024, The Authors. All rights reserved.

关键词： computer aided diagnosis

Robust voxelization and visualization by improved tetrahedral mesh generation

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Chen, Joseph Tai, Ko-Wei Chen, Wen-Chin Ouhyoung, Ming Graduate Institute of Networking and Multimedia National Taiwan University Department of Computer Science and Information Engineering National Taiwan University

When obtaining interior 3D voxel data from triangular meshes, most existing methods fail to handle low quality meshes which happens to take up a big portion on the internet. In this work we present a robust voxelization method that is based on tetrahedral mesh generation within a user defined error bound. Comparing to other tetrahedral mesh generation methods, our method produces much higher quality tetrahedral meshes as the intermediate outcome, which allows us to utilize a faster voxelization algorithm that is based on a stronger assumption. We show the results comparing to various methods including the state-of-the-art. Our contribution includes a framework which takes triangular mesh as an input and produces voxelized data, a proof to an unproved algorithm that performs better than the state-of-the-art, and various experiments including parallelization built on the GPU and CPU. We further tested our method on various dataset including Princeton ModelNet and Thingi10k to show the robustness of the framework, where near 100% availability is achieved, while others can only achieve around 50%. © 2021, CC BY.

关键词：

Unsupervised Training Framework for 3D Point Cloud Object Detection Model 25

学校读者我要写书评

暂无评论

Unsupervised Training Framework for 3D Point Cloud Object De...

Proceedings of the 40th ACM/SIGAPP Symposium on Applied Computing

作者： Min-Jyun Wu Hsiang-Jui Lin Shih-Hao Hung Chi-Sheng Shih Dept. of Computer Science and Information Engineering National Taiwan University Taipei Taiwan High Performance and Scientific Computing (HPSC) Center National Taiwan University Taipei Taiwan Graduate Institute of Networking and Multimedia National Taiwan University Taipei Taiwan Center of Artificial Intelligence and Advanced Robotics National Taiwan University Taipei Taiwan

ISBN: (纸本)9798400706295

Point cloud data frames are critical, if not indispensable, for precise robot navigation and localization, but training the object detection models for them remains challenging. Many models require labeled 3D objects to train the model. However, the sparse and occluded 3D point cloud data make it difficult, if not impossible, to automate the labeling process. This work proposes a training framework to generate 3D labels on point clouds to tackle the aforementioned challenges. The proposed method takes advantage of the consecutive presence of the same object on different frames to automate the labeling process. The experimental results show that the unsupervised framework trains a robust model for 3D object detection. On the roadside data, the model archives 90.27% AP for scooters and 91.33% AP for cars. On nuScenes dataset, the framework demonstrates the detection precision doubles on IoU 0.25 and IoU 0.5 when the recalls remain similar, compared to the model trained by the MODEST framework.

关键词： computer systems organization

An Automatic Meter Reading System Using computer Vision Technology and Hydroelectricity Auxiliary Power Supply 11

学校读者我要写书评

暂无评论

An Automatic Meter Reading System Using Computer Vision Tech...

11th IEEE Global Conference on Consumer Electronics, GCCE 2022

作者： Wang, Dao-Jiun Fanjiang, Yong-Yi Hung, Chi-Huang Yeh, Yu-Hsiang Hsiao, Yao-Yun Hsiao, Sheng-En Graduate Institute of Applied Science and Engineering Fu Jen Catholic University New Taipei City Taiwan Graduate Institute of Applied Science and Engineering Dept. of Computer Science and Information Engineering Fu Jen Catholic University New Taipei City Taiwan Dept. of Information Technology Graduate Institute of Applied Science and Engineering Fu Jen Catholic University Lee-Ming Institute of Technology New Taipei City Taiwan Fu Jen Catholic University Department of Computer Science and Information Engineering New Taipei City Taiwan Lee-Ming Institute of Technology Dept. of Digital Multimedia New Taipei City Taiwan Lee-Ming Institute of Technology Dept. of Fashion Business Management New Taipei City Taiwan

ISBN: (纸本)9781665492324

Image recognition based automatic meter reading (AMR) method is getting popular nowadays. As traditional mechanical water meters are usually placed in poor environments, if we install the AMR on the water meter, the power supply of the AMR is an issue that needs to be considered thoughtfully. This study proposes to place a hydroelectric generator in the water pipe to provide auxiliary power in order to improve the power supply of AMR system. The system can also automatically identify the value of the water meter through the cloud service, so as to reduce the personnel cost of the water company, and further manage and control the hydraulic resources more conveniently. © 2022 IEEE.

关键词： Character recognition

Self-Supervised Multi-LiDAR Object View Generation Using Single LiDAR

学校读者我要写书评

暂无评论

Self-Supervised Multi-LiDAR Object View Generation Using Sin...

IEEE International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA)

作者： Yi-Hung Kuo Chi-Sheng Shih Hsiang-Jui Lin Shih-Hao Hung Department of Computer Science and Information Engineering National Taiwan University Taipei Taiwan Graduate Institute of Networking and Multimedia National Taiwan University Taipei Taiwan High Performance and Scientific Computing Center National Taiwan University Taipei Taiwan Mohamed bin Zayed University of Artificial Intelligence Abu Dhabi

3D object detection using point clouds has been widely used for self-driving vehicles, roadside vehicle detection, and tracking. However, the detection accuracy using single-LiDAR-scanned point clouds suffers from occlusion to have accurate bounding boxes. Several previous works proposed compensating for the point cloud to improve 3D detection. This work designed a self-labeling method that exploits the characteristics of multiple LiDAR-scanned point clouds collected on roadside units. This labeling method allows the point completion model to be trained without human-labeled bounding boxes. The point completion model uses the single-LiDAR-scanned point cloud as input and uses the extracted object points from the fused point cloud as the training target. The voxel IoU and recall of the point-completed point clouds increased by 17% and 22% compared to that of the raw point clouds. Additionally, the detection performance of DBSCAN, i.e., 3D AP@IoU=0.25 increased by 20% after applying the point completion algorithm.

关键词：

FAD-SAR: A Novel Fishing Activity Detection System via Synthetic Aperture Radar Images Based on Deep Learning Method

学校读者我要写书评

暂无评论

FAD-SAR: A Novel Fishing Activity Detection System via Synth...

IEEE International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS)

作者： Yanbing Bai Siao Li Rui-Yang Ju Zihao Yang Jinze Yu Jen-Shiun Chiang Center for Applied Statistics School of Statistics Renmin University of China Beijing China Graduate Institute of Networking and Multimedia National Taiwan University Taipei City Taiwan Department of Mechanical Systems Engineering Tokyo University of Agriculture and Technology Tokyo Japan Department of Electrical and Computer Engineering Tamkang University New Taipei City Taiwan

ISBN: (数字)9798350389210

ISBN: (纸本)9798350389227

Illegal, unreported, and unregulated (IUU) fishing activities seriously affect various aspects of human life. However, traditional methods for detecting and monitoring IUU fishing activities at sea have limitations. Although synthetic aperture radar (SAR) can complement existing vessel detection systems, extracting useful information from SAR images using traditional methods remains a challenge, especially in IUU fishing. This paper proposes a deep learning based fishing activity detection system, which is implemented on the xView3 dataset using six classical object detection models: SSD, RetinaNet, FSAF, FCOS, Faster R-CNN, and Cascade R-CNN. In addition, this work employs different enhancement techniques to improve the performance of the Faster R-CNN model. The experimental results demonstrate that training the Faster R-CNN model using the Online Hard Example Mining (OHEM) strategy increases the Avg-F1 value from 0.212 to 0.216.

关键词： Deep learning Training Radar cross-sections Object detection Signal processing Radar polarimetry Data mining Synthetic aperture radar Remote sensing Monitoring

Automated whole-breast ultrasound tumor diagnosis using attention-inception network

学校读者我要写书评

暂无评论

Physica Medica 2025年 134卷 104989-104989页

作者： Zhang, Jun Huang, Yao-Sian Wang, You-Wei Xiang, Huiling Lin, Xi Chang, Ruey-Feng Graduate Institute of Biomedical Electronics and Bioinformatics National Taiwan University Taipei Taiwan Department of Computer Science and Information Engineering National Changhua University of Education Changhua Taiwan Department of Computer Science and Information Engineering National Taiwan University Taipei Taiwan Department of Ultrasound Sun Yat-sen University Cancer Center Guangzhou China Graduate Institute of Network and Multimedia National Taiwan University Taipei Taiwan

Purpose: Automated Whole-Breast Ultrasound (ABUS) has been widely used as an important tool in breast cancer diagnosis due to the ability of this technique to provide complete three-dimensional (3D) images of breasts. To eliminate the risk of misdiagnosis, computer-aided diagnosis (CADx) systems have been proposed to assist radiologists. Convolutional neural networks (CNNs), renowned for the automatic feature extraction capabilities, have developed rapidly in medical image analysis, and this study proposes a CADx system based on 3D CNN for ABUS. Materials and methods: This study used a private dataset collected at Sun Yat-Sen University Cancer Center (SYSUCC) from 396 breast tumor patients. First, the tumor volume of interest (VOI) was extracted and resized, and then the tumor was enhanced by histogram equalization. Second, a 3D U-Net++ was employed to segment the tumor mask. Finally, the VOI, the enhanced VOI, and the corresponding tumor mask were fed into a 3D Attention-Inception network to classify the tumor as benign or malignant. Results: The experiment results indicate an accuracy of 89.4%, a sensitivity of 91.2%, a specificity of 87.6%, and an area under the receiver operating characteristic curve (AUC) of 0.9262, which suggests that the proposed CADx system for ABUS images rivals the performance of experienced radiologists in tumor diagnosis tasks. Conclusion: This study proposes a CADx system consisting of a 3D U-Net++ tumor segmentation model and a 3D attention inception neural network tumor classification model for diagnosis in ABUS images. The results indicate that the proposed CADx system is effective and efficient in tumor diagnosis tasks. © 2025 Associazione Italiana di Fisica Medica e Sanitaria

关键词： ABUS Attention mechanism Breast cancer CADx Deep learning Group convolution

YOLOv8-ResCBAM: YOLOv8 Based on An Effective Attention Module for Pediatric Wrist Fracture Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ju, Rui-Yang Chien, Chun-Tse Chiang, Jen-Shiun Graduate Institute of Networking and Multimedia National Taiwan University No. 1 Sec. 4 Roosevelt Rd. Taipei City106319 Taiwan Department of Electrical and Computer Engineering Tamkang University No.151 Yingzhuan Rd. Tamsui Dist. New Taipei City251301 Taiwan

Wrist trauma and even fractures occur frequently in daily life, particularly among children who account for a significant proportion of fracture cases. Before performing surgery, surgeons often request patients to undergo X-ray imaging first, and prepare for the surgery based on the analysis of the X-ray images. With the development of neural networks, You Only Look Once (YOLO) series models have been widely used in fracture detection for computer-Assisted Diagnosis, where the YOLOv8 model has obtained the satisfactory results. Applying the attention modules to neural networks is one of the effective methods to improve the model performance. This paper proposes YOLOv8-ResCBAM, which incorporates Convolutional Block Attention Module integrated with resblock (ResCBAM) into the original YOLOv8 network architecture. The experimental results on the GRAZPEDWRI-DX dataset demonstrate that the mean Average Precision calculated at Intersection over Union threshold of 0.5 (mAP 50) of the proposed model increased from 63.6% of the original YOLOv8 model to 65.8%, which achieves the state-of-the-art performance. The implementation code is available at https://***/RuiyangJu/Fracture_Detection_Improved_YOLOv8. Copyright © 2024, The Authors. All rights reserved.

关键词： Medical image processing