检索结果-内蒙古大学图书馆

ORF-Net: deep Omni-Supervised Rib Fracture Detection from Chest CT Scans 1

25th International Conference on Medical image Computing and Computer Assisted Intervention (MICCAI)

作者： Chai, Zhizhong Lin, Huangjing Luo, Luyang Heng, Pheng-Ann Chen, Hao Imsight Res Lab Shenzhen Peoples R China Chinese Univ Hong Kong Dept Comp Sci & Engn Hong Kong Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Guangdong Prov Key Lab Comp Vision & Virtual Real Shenzhen Peoples R China Hong Kong Univ Sci & Technol Dept Comp Sci & Engn Hong Kong Peoples R China

ISBN: (数字)9783031164378

ISBN: (纸本)9783031164378;9783031164361

Most of the existing object detection works are based on the bounding box annotation: each object has a precise annotated box. However, for rib fractures, the bounding box annotation is very labor-intensive and time-consuming because radiologists need to investigate and annotate the rib fractures on a slice-by-slice basis. Although a few studies have proposed weakly-supervised methods or semi-supervised methods, they could not handle different forms of supervision simultaneously. In this paper, we proposed a novel omni-supervised object detection network, which can exploit multiple different forms of annotated data to further improve the detection performance. Specifically, the proposed network contains an omni-supervised detection head, in which each form of annotation data corresponds to a unique classification branch. Furthermore, we proposed a dynamic label assignment strategy for different annotated forms of data to facilitate better learning for each branch. Moreover, we also design a confidence-aware classification loss to emphasize the samples with high confidence and further improve the model's performance. Extensive experiments conducted on the testing dataset show our proposed method outperforms other state-of-the-art approaches consistently, demonstrating the efficacy of deep omni-supervised learning on improving rib fracture detection performance.

关键词： Omni-supervised learning Rib fracture Object detection

来源：评论

学校读者我要写书评

暂无评论

Run Your 3D Object Detector on NVIDIA Jetson Platforms:A Benchmark Analysis

引用

SENSORS 2023年第8期23卷 4005-4005页

作者： Choe, Chungjae Choe, Minjae Jung, Sungwook Korea Elect Technol Inst Autonomous IoT Res Ctr Seongnam 13509 South Korea Caterpillar Inc Peoria IL 61629 USA

This paper presents a benchmark analysis of NVIDIA Jetson platforms when operating deep learning-based 3D object detection frameworks. Three-dimensional (3D) object detection could be highly beneficial for the autonomous navigation of robotic platforms, such as autonomous vehicles, robots, and drones. Since the function provides one-shot inference that extracts 3D positions with depth information and the heading direction of neighboring objects, robots can generate a reliable path to navigate without collision. To enable the smooth functioning of 3D object detection, several approaches have been developed to build detectors using deep learning for fast and accurate inference. In this paper, we investigate 3D object detectors and analyze their performance on the NVIDIA Jetson series that contain an onboard graphical processing unit (GPU) for deep learning computation. Since robotic platforms often require real-time control to avoid dynamic obstacles, onboard processing with a built-in computer is an emerging trend. The Jetson series satisfies such requirements with a compact board size and suitable computational performance for autonomous navigation. However, a proper benchmark that analyzes the Jetson for a computationally expensive task, such as point cloud processing, has not yet been extensively studied. In order to examine the Jetson series for such expensive tasks, we tested the performance of all commercially available boards (i.e., Nano, TX2, NX, and AGX) with state-of-the-art 3D object detectors. We also evaluated the effect of the TensorRT library to optimize a deep learning model for faster inference and lower resource utilization on the Jetson platforms. We present benchmark results in terms of three metrics, including detection accuracy, frame per second (FPS), and resource usage with power consumption. From the experiments, we observe that all Jetson boards, on average, consume over 80% of GPU resources. Moreover, TensorRT could remarkably increase infe

关键词： 3D object detection deep learning Jetson platforms benchmark

来源：评论

学校读者我要写书评

暂无评论

Landslide Susceptibility for Communities Based on Satellite images Using deep learning Algorithms

Landslide Susceptibility for Communities Based on Satellite ...

引用

International Conference on Intelligent Systems and Sustainable Computing, ICISSC 2021

作者： Sridharan, Aadityan Ajai, A. S. Remya Gopalan, Sundararaman Department of Physics Amrita Vishwa Vidyapeetham Amritapuri India Department of Electronics and Communication Engineering Amrita Vishwa Vidyapeetham Amritapuri India

ISBN: (纸本)9789811900105

High resolution satellite images serve as an eye in the sky and are helpful in generating landslide inventories. Satellite images of landslides can be used as an input for deep learning models that are specifically designed for image processing and assess the risk posed by landslides on communities. Fully satellite image-based landslide risk evaluation is still a new venture that needs a lot of improvements before it can be deployed in real-time. In this work, we assess landslide hazard to communities near debris scars of slopes that have experienced landslides in the past, using three pre-trained deep learning algorithms. We followed a 70:30 scheme for training and validation with over 2000 images of two classes: Urban and Debris. Finally, the trained neural network is tested on images that contain human settlements near the debris scars. Based on class probabilities predicted by the algorithm, the sites were ranked for eminent risks in the future. It was observed that the validation accuracy of Alexnet, Resnet and NASNet-Large were 92%, 96% and 98%, respectively. The risk classification for four test images indicates: Alexnet and NASNet-Large extracted features same as observed in the satellite images however, Resnet50 is very sensitive to the features and could not predict the same as observed in satellite images. Given the accuracy of predictions, such algorithms can be further modified and deployed to create landslide hazard risk maps for various communities around the world. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Landslides

来源：评论

学校读者我要写书评

暂无评论

Greenguard: CNN-based system for Intelligent Plant Disease Classification

Greenguard: CNN-based system for Intelligent Plant Disease C...

引用

International Conference on Inventive Systems and Control (ICISC)

作者： Sk Sajida Sultana Syed Shareefunnisa Verella Sai Spandana Annapureddy Shiva Bhargavi Gunji Sailaja Chamakura Rechal Reddy Dept of CSE Vignan’s Foundation for Science Technology & Research Vadlamudi Guntur AP India

ISBN: (数字)9798350386578

ISBN: (纸本)9798350386585

Plant diseases significantly impact agricultural output, yet traditional detection methods are costly, time-consuming, and inaccurate. This study proposes a more efficient disease recognition model based on image processing and deep learning, specifically Convolutional Neural Networks (CNNs). These networks perform well at analyzing leaf images to identify plant diseases. However, existing models trained on ideal conditions struggle with real-world field photos featuring complex backgrounds and multiple leaves. To address this challenge, the study will test and compare various CNN architectures like VGG19, AlexNet, and ResNet aiming to find the model with the highest accuracy in detecting diseases under realistic conditions. The ultimate goal is to develop a tool that empowers farmers to quickly and accurately identify crop diseases, ultimately leading to improved agricultural output and reduced losses.

关键词： deep learning Plant diseases Accuracy image recognition Refining Process control Crops

来源：评论

学校读者我要写书评

暂无评论

An Improved RT-DETR Model for UAV-Based Power Line Inspection Under Various Weather Conditions

An Improved RT-DETR Model for UAV-Based Power Line Inspectio...

引用

Chinese Automation Congress (CAC)

作者： Hantao Ding Chen Zhou Cheng Lian School of Automation Wuhan University of Technology WuHan China Huizhou Power Supply Bureau of Guangdong Power Gria Huizhou China

ISBN: (数字)9798350368604

ISBN: (纸本)9798350368611

With the rapid advancement of deep learning, convolutional neural network-based target detection models have been effectively applied in UAV power inspection. However, the success of convolutional neural networks (CNNs) heavily relies on image augmentation and post-processing, both of which consume significant amounts of time in real-time object detection. This paper presents an end-to-end improved RT-DETR model that eliminates the post-processing module and introduces a reweighted L1 Loss to enhance the model's ability to learn small objects. Additionally, to improve the model's multi-scale feature fusion capability, we incorporate the GELAN module, combining CSPNet and ELAN models, allowing the network to more effectively propagate and aggregate feature information from different levels. Furthermore, to solve the shortage of power inspection datasets, we propose a Multi-Weather UAV Power Inspection dataset (MWUPI). Experimental results demonstrate that the improved RT-DETR model achieves high accuracy on this dataset while maintaining the fast detection speed of the original RT-DETR, meeting the performance requirements of real-time object detection. The codes for this study are publicly available at https://***/HantaoDing/Improved-RT-DETR-for-Power-Inspection.

关键词： Accuracy Semantics Object detection Inspection Autonomous aerial vehicles Insulators real-time systems image augmentation Convolutional neural networks Meteorology

来源：评论

学校读者我要写书评

暂无评论

Efficient Data Extraction from Handwritten Forms: A Structured Pipeline Solution

Efficient Data Extraction from Handwritten Forms: A Structur...

引用

Artificial Intelligence Trends and Pattern Recognition (ICAITPR), International Conference on

作者： Shivani Patel Krisha Borana Neha Katre Vinaya Sawant Harshal Dalvi Department of Information Technology D.J. Sanghvi College Of Engineering Mumbai India

ISBN: (数字)9798350368000

ISBN: (纸本)9798350368017

This paper presents an approach to systematically extract data from handwritten forms and convert them to structured data. Unlike traditional handwriting character recognition systems, proposed approach integrates techniques such as alpha blending for precise image alignment and deep learning-based models for enhanced text recognition accuracy. The pipeline begins with the automated detection and delineation of form fields (like Name, DOB, Occupation) followed by sophisticated pre-processing steps that include noise reduction and threshold management to ensure optimal image quality. The system then employs the proposed trained model to recognize handwritten text and match it to the extracted entities like key value pairs. The values are then appended to structured data. The proposed method has been evaluated, demonstrating accuracy and reliability in diverse real-world scenarios. This method would bypass the need to individually extract values and handwritten texts, saving time and effort.

关键词： deep learning Training Accuracy Text recognition Pipelines Optical character recognition Robustness Data mining Character recognition Research and development

来源：评论

学校读者我要写书评

暂无评论

Hybrid Contour and Distance-based Loss for Enhanced Bladder and Prostate Segmentation in CT images

Hybrid Contour and Distance-based Loss for Enhanced Bladder ...

引用

2025 Conference on Medical Imaging

作者： Ribeiro, Raul Ferrete Torresa, Helena R. Moane, Siobhan Morais, Pedro Vilaca, Joao L. IPCA 2Ai Sch Technol Barcelos Portugal TUS LIFE Res Inst Dept Appl Sci Limerick Ireland LASI Associate Lab Intelligent Syst Guimaraes Portugal

ISBN: (纸本)9781510685901

Segmentation of bladder and prostate in Computed Tomography (CT) scans is an important task in clinical practice for aiding in diagnosis, monitoring conditions, and identifying cancerous lesions, which affect around 2.1 million patients suffering from lower urinary tract symptoms. Traditionally, radiologists manually segment these structures on CT scans to identify and locate unusual tissues, shapes, and worrying areas. Despite its importance, it has been considered a backbreaking task due to: i) difficulty, ii) time-consuming, and iii) subject to inter- and intra-observer variability. This research proposes a novel approach for delineating the bladder and prostate in CT scans. For that, a new hybrid loss function was implemented to enhance bladder and prostate segmentation, by gathering the information of the segmentation, contour, and distance map. Moreover, this study also includes a comparative analysis of DL architectures to assess the improvement of the segmentation detail. The proposed method showed strong performance on the AMOS22 and PAEC datasets, containing 292 training, 82 validation, and 43 testing images. For the AMOS22 dataset, mean average surface distance (aSSD) for the bladder and prostate were 1.5 +/- 1.54 mm and 1.65 +/- 1.78 mm, respectively, and the 95th percentile of Hausdorff Distance (HD95) values were 7.31 +/- 11.34 mm and 5.04 +/- 3.42 mm, respectively, corresponding to mean Dice scores of 88.76 +/- 9.02% and 85.90 +/- 12.93%. The method on the PAEC dataset yielded mean aSSDs of 1.73 +/- 1.35 mm and 3.26 +/- 1.67 mm with reductions of 66.47% and 36.08% for the bladder and prostate, respectively, with average Dice scores of 90.80 +/- 7.91% and 80.57 +/- 7.07%, and average HD95 values of 7.82 +/- 7.80 mm and 8.67 +/- 3.18 mm. The obtained results suggest the potential clinical interest of this AI model to assist doctors in performing screenings of the bladder and prostate.

关键词： deep learning Segmentation Distance Map Discrepancy Bladder Prostate AMOS PAEC

来源：评论

学校读者我要写书评

暂无评论

A Low Spectral Bias Generative Adversarial Model for image Generation 8th

A Low Spectral Bias Generative Adversarial Model for Image G...

引用

8th International Conference on Pioneering Computer Scientists, Engineers and Educators (ICPCSEE)

作者： Xu, Lei Liu, Zhentao Liu, Peng Cai, Liyan Peng Cheng Lab Shenzhen Peoples R China Sun Yat Sen Univ Guangzhou Peoples R China

ISBN: (纸本)9789811951947;9789811951930

We propose a systematic analysis of the neglected spectral bias in the frequency domain in this paper. Traditional generative adversarial networks (GANs) try to fulfill the details of images by designing specific network architectures or losses, focusing on generating visually qualitative images. The convolution theorem shows that image processing in the frequency domain is parallelizable and performs better and faster than that in the spatial domain. However, there is little work about discussing the bias of frequency features between the generated images and the real ones. In this paper, we first empirically demonstrate the general distribution bias across datasets and GANs with different sampling methods. Then, we explain the causes of the spectral bias through the deduction that reconsiders the sampling process of the GAN generator. Based on these studies, we provide a low-spectral-bias hybrid generative model to reduce the spectral bias and improve the quality of the generated images.

关键词： deep learning applications image generation models Generative adversarial network

来源：评论

学校读者我要写书评

暂无评论

Machine learning for yield prediction in Fergana valley, Central Asia

引用

Journal of the Saudi Society of Agricultural Sciences 2023年第2期22卷 107-120页

作者： Singh Boori, Mukesh Choudhary, Komal Paringer, Rustam Kupriyanov, Alexander Scientific Research Laboratory of Automated Systems of Scientific Research (SRL-35) Samara National Research University Samara Russian Federation Department of Land Surveying and Geo- Informatics Smart Cities Research Institute The Hong Kong Polytechnic University Kowloon Hong Kong Image Processing Systems Institute of the RAS–Branch of the FSRC “Crystallography and Photonics” Samara Russian Federation

Accurate yield prediction is essential for growers, researchers, governments, the farming industry, and policymakers for social peace, food safety, security, and sustainable development. The results of earlier techniques of data collecting and analysis for yield forecasts were typically delayed, expensive, time-consuming, site-specific, and riddled with errors and uncertainties. This study is a novel approach to using high-resolution satellite data in conjunction with environmental and topographic data to predict wheat yield variability at the farm scale using machine learning. In this research, winter wheat yield prediction was based on 36 indicators in machine learning using correlation and different regression models. Winter wheat yield was predicted using linear regression (LR), decision tree (DT), and random forest (RF) regression models with scikit-learn in machine learning. More than 10,000 data points from 45 farms were trained and validated in Fergana valley, Central Asia. Results indicate that at 10 m resolution using Sentinel-2 and other secondary data such as topographic, soil, environmental, and filed data can generate an accurate wheat yield prediction map. The accuracy of all regressions is lowest for LR (R²:95, RMSE: 2.31), highest for RF (R²:98, RMSE: 1.40), and intermediate for DT regression (R²:97, RMSE: 1.85). Results also indicate that prediction in the early stage of the crop is less accurate in comparison to harvesting time as LR (R²:85, RMSE: 2.66), DT (R²:95, RMSE: 2.06), RF (R²:97, RMSE: 1.54) have different R² and RMSE values. Appling the RF model, the winter wheat prediction is 3.29 to 4.30 t/ha therefore the total wheat production is approximately 100 t in the study area. Thus this study will demonstrate the capability of high-resolution satellite imagery and secondary data for highly accurate real-time crop yield prediction at the field scale, which can be used

关键词： Food security Machine learning Phenology Precision agriculture Regression Sentinel-2 Spectral-indices Yield prediction

来源：评论

学校读者我要写书评

暂无评论

Go-selfies: A fast selfies background removal method using ResU-net deep learning 28

Go-selfies: A fast selfies background removal method using R...

引用

28th European Signal processing Conference, EUSIPCO 2020

作者： Wu, Yunan Department of Biomedical Engineering Northwestern University EvanstonIL United States

ISBN: (纸本)9789082797053

The selfies play an important role in recording meaningful moment in human's daily life. In most cases, before sharing photos, people often synthesis attractive images on some phone applications, such as Photoshop. While these kinds of software have reached good performance nowadays, they are too complex for simple life usage. In this work, we proposed an automatic segmentation model unique to segment human selfies photos. We first constructed a large photo segmentation database and built 8 different models based on resolution, image size and whether or not to use transfer learning and picked the best one among them. We then applied cyclical learning rate method and pre-trained encoder network to fine tune our models. Finally, our best model tested on Google images demonstrated satisfying promising results on both accuracy scores and losses, which will be the precondition in real-time segmentation. We named this lovely web product as "Go Selfies". © 2021 European Signal processing Conference, EUSIPCO. All rights reserved.

关键词： deep learning

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：