检索结果-内蒙古大学图书馆

3rd International Symposium on Intelligent Unmanned Systems and Artificial Intelligence (SIUSAI)

作者： Feng, Zhenzhen Guo, Xiaoyong Liu, Yuntao Cui, Can Tianjin Univ Sci & Technol Coll Elect Informat & Automat Tianjin Peoples R China Ocean Univ China Coll Engn Qingdao Peoples R China

ISBN: (纸本)9798400710025

In this paper, an edge computing box design method based on deep learning and embedded system is proposed to address the problems of complex harbor operating environment and frequent safety accidents. The computing box is based on Nvidia's Jetson Nano hardware architecture, and a Pedestrian net deep neural network model specialized in human and vehicle detection is developed through TensorRT and deepstream. The computational box designed by this method supports secondary development and fast iteration, and can be flexibly adapted to various mainstream video processing AI models. Experiments show that the computational box has good computational real-time performance and extremely fast inference speed, which guarantees the deploy ability of port security and intelligent supervision in embedded terminal equipment.

关键词： Intelligent supervision Edge computing Jetson nano deep neural networks Human-vehicle detection

来源：评论

学校读者我要写书评

暂无评论

An AI-Based Low-Risk Lung Health image Visualization Framework Using LR-ULDCT

引用

JOURNAL OF IMAGING INFORMATICS IN MEDICINE 2024年第5期37卷 2047-2062页

作者： Rai, Swati Bhatt, Jignesh S. Patra, Sarat Kumar Indian Inst Informat Technol Vadodara Vadodara India Natl Inst Technol Agartala Agartala India

In this article, we propose an AI-based low-risk visualization framework for lung health monitoring using low-resolution ultra-low-dose CT (LR-ULDCT). We present a novel deep cascade processing workflow to achieve diagnostic visualization on LR-ULDCT (<0.3 mSv) at par high-resolution CT (HRCT) of 100 mSV radiation technology. To this end, we build a low-risk and affordable deep cascade network comprising three sequential deep processes: restoration, super-resolution (SR), and segmentation. Given degraded LR-ULDCT, the first novel network unsupervisedly learns restoration function from augmenting patch-based dictionaries and residuals. The restored version is then super-resolved (SR) for target (sensor) resolution. Here, we combine perceptual and adversarial losses in novel GAN to establish the closeness between probability distributions of generated SR-ULDCT and restored LR-ULDCT. Thus SR-ULDCT is presented to the segmentation network that first separates the chest portion from SR-ULDCT followed by lobe-wise colorization. Finally, we extract five lobes to account for the presence of ground glass opacity (GGO) in the lung. Hence, our AI-based system provides low-risk visualization of input degraded LR-ULDCT to various stages, i.e., restored LR-ULDCT, restored SR-ULDCT, and segmented SR-ULDCT, and achieves diagnostic power of HRCT. We perform case studies by experimenting on real datasets of COVID-19, pneumonia, and pulmonary edema/congestion while comparing our results with state-of-the-art. Ablation experiments are conducted for better visualizing different operating pipelines. Finally, we present a verification report by fourteen (14) experienced radiologists and pulmonologists.

关键词： Artificial intelligence deep learning Lung infection Reconstruction Visualization system Ultra-low-dose computed tomography

来源：评论

学校读者我要写书评

暂无评论

Identification of Disaster Related Tweets for real-time Response using Machine learning

Identification of Disaster Related Tweets for Real-Time Resp...

引用

2024 International Conference on Recent Innovation in Smart and Sustainable Technology, ICRISST 2024

作者： Dagar, Ishant Gusain, Raj Shekhar, Sushant Vidyarthi, Anurag Prakash, Rishi Gowri, R. Deemed to Be University Propagation Research Lab Department of Electronics & Communication Engineering Graphic Era Dehradun India Graphic Era Hill University Department of Computer Science Engineering Dehradun India

ISBN: (纸本)9798350309140

This paper presents the identification system for social media tweets about the disaster. A sophisticated machine learning framework is designed to classify tweets as disaster-related or not. In an age where social media platforms serve as real-time information provider during critical events, the ability to identify disaster-related content swiftly and accurately is necessary. Using deep learning techniques for natural language processing, this system represents a significant advancement in disaster identification through social media. The dataset used for training and evaluating the model is taken from Kaggle. This dataset contains a variety of labeled tweets, including essential information such as keywords, locations, and the tweet content itself. Robust data preprocessing techniques are employed to handle missing values, encode categorical variables, and examine the class distribution. These steps ensure the dataset's compatibility for model training. The core of the Disaster Identification System is in its neural network architecture. It incorporates bidirectional Long Short-Term Memory (LSTM) layers, spatial dropout, and dense layers to effectively capture content related information from the tweet content. The model is trained using binary cross-entropy loss and the Adam optimizer, which is a powerful predictive engine. Furthermore, the Disaster Identification System has a real-time prediction function that enables users to input tweets and receive instant assessments of whether the content is disaster-related. This feature has the potential to increase the speed and accuracy of information dissemination during critical events. The Disaster Identification System represents a step for disaster prediction and real-time information filtering. Its application extends beyond disaster management to cover emergency response, public safety, and crisis communication. This paper provides an in-depth exploration of the system's architecture and capabilities, providing

关键词： Tweets

来源：评论

学校读者我要写书评

暂无评论

deep learning-Based Fruit and Vegetable Detection 9

Deep Learning-Based Fruit and Vegetable Detection

引用

9th International Conference on Information Technology and Digital Applications, ICITDA 2024

作者： Kang, Ho Jun Yogarayan, Sumendra Mogan, Jashila Nair Min, Pa Pa Faculty of Information Science and Technology Melaka Malaysia

ISBN: (纸本)9798331510732

In recent times, suppliers in the agricultural sector are under increasing pressure to enhance efficiency in sorting and categorizing fruits and vegetables. Ensuring accurate and efficient detection is crucial for optimizing supply chain operations. This study presents the development of an embedded system using deep learning techniques to address these challenges. Specifically, a Convolutional Neural Network (CNN) model is implemented on a Raspberry Pi 5, designed to improve the accuracy of fruit and vegetable identification for suppliers. The system integrates image capture and real-time detection capabilities, streamlining the sorting process. The dataset, which includes images of various produce, undergoes preprocessing steps such as grayscale conversion and feature scaling. A train-test split is employed, with 60% of the data allocated for training, 20% for validation, and 20% for testing. The CNN model is trained and evaluated, demonstrating high classification accuracy. © 2024 IEEE.

关键词： Fruits

来源：评论

学校读者我要写书评

暂无评论

real-time Traffic Intensity Estimation and Management using deep learning Techniques 2

Real-Time Traffic Intensity Estimation and Management using ...

引用

2nd IEEE International Conference on Networks, Multimedia and Information Technology, NMITCON 2024

作者： Murhekar, Suryakant Kamble, Yash Dongre, Snehlata Dasarwar, Priya Khan, Tabassum H Patil, Vaishali G H Raisoni College Of Engineering Nagpur India Dept. of Computer Science and Engineering G H Raisoni College Of Engineering Pune India Dept. of Artificial Intelligence G H Raisoni Institute Of Engineering and Technology Nagpur India

ISBN: (纸本)9798350372892

In India, traffic flow estimation is crucial for effective transportation management and urban planning. This research proposes a comprehensive approach leveraging deep learning techniques, particularly the YOLOv8 object detection model, to estimate road traffic flow in urban environments. The model is trained on the widely used COCO dataset, ensuring robust performance in detecting various objects, including vehicles, pedestrians, and cyclists. Subsequently, the model is fine-tuned on a custom dataset comprising annotated traffic images, enabling it to adapt specifically to the complexities of urban traffic scenarios. The research methodology encompasses several key steps. Firstly, the pretrained YOLOv8 model is utilized to perform inference on sample images, demonstrating its capability to accurately detect vehicles within the urban environment. Visualization and analysis of the detection results showcase the model's effectiveness in identifying vehicles with high confidence levels, laying a foundation for further experimentation. Following the initial assessment, the model undergoes fine-tuning on the custom dataset, optimizing its performance for the task of traffic flow estimation. Training parameters such as epochs, image size, and learning rate are carefully tuned to achieve optimal results. Additionally, metrics such as box loss, classification loss, and distribution focal loss are monitored during training to assess model convergence and performance. Upon completing the training phase, the model's performance is evaluated using a validation set, providing insights into its generalization ability and effectiveness in real-world scenarios. The evaluation metrics, including precision, recall, and F1 score, offer quantitative measures of the model's performance, further validating its utility for traffic flow estimation tasks. Furthermore, the research extends beyond static image analysis to real-time traffic intensity estimation using video data. By integratin

关键词： Urban transportation

来源：评论

学校读者我要写书评

暂无评论

Segmentation of kidney stones in endoscopic video feeds

Segmentation of kidney stones in endoscopic video feeds

引用

Conference on Medical Imaging - image processing

作者： Stoebner, Zachary A. Lu, Daiwei Hong, Seok Hee Kavoussi, Nicholas L. Oguz, Ipek Vanderbilt Univ Nashville TN 37235 USA Vanderbilt Univ Med Ctr Nashville TN USA

ISBN: (数字)9781510649408

ISBN: (纸本)9781510649408;9781510649392

image segmentation has been increasingly applied in medical settings as recent developments have skyrocketed the potential applications of deep learning. Urology, specifically, is one field of medicine that is primed for the adoption of a real-time image segmentation system with the long-term aim of automating endoscopic stone treatment. In this project, we explored supervised deep learning models to annotate kidney stones in surgical endoscopic video feeds. In this paper, we describe how we built a dataset from the raw videos and how we developed a pipeline to automate as much of the process as possible. For the segmentation task, we adapted and analyzed three baseline deep learning models - U-Net, U-Net++, and DenseNet - to predict annotations on the frames of the endoscopic videos with the highest accuracy above 90%. To show clinical potential for real-time use, we also confirmed that our best trained model can accurately annotate new videos at 30 frames per second. Our results demonstrate that the proposed method justifies continued development and study of image segmentation to annotate ureteroscopic video feeds.

关键词： segmentation deep learning endoscopy computer vision

来源：评论

学校读者我要写书评

暂无评论

Enhancing Prosthetic Control with Ultrasound images: A Convolutional Neural Network Approach for Hand Gesture Recognition 4

Enhancing Prosthetic Control with Ultrasound Images: A Convo...

引用

4th Modeling, Estimation, and Control Conference (MECC)

作者： Chen, Yun Bao, Xuefeng He, Hongsheng Zhang, Qiang Univ Alabama Dept Mech Engn Tuscaloosa AL 35487 USA Univ Wisconsin Dept Biomed Engn Milwaukee WI 53045 USA Univ Alabama Dept Comp Sci Tuscaloosa AL 35487 USA

Human hand gesture recognition using biological signals from the forearm is an increasingly significant area of research, with implications across various fields such as prosthetic development, rehabilitation, and human-machine interaction. However, traditional hand gesture recognition with surface electromyography (sEMG) technique has some challenges, including cross-talk from neighboring muscles, low signal-to-noise ratio, and inability to measure deep muscles. In the current study, we proposed to use brightness mode (B-mode) ultrasound images from the muscles of the forearm anterior side as an alternative neuromuscular interface to recognize hand gestures. We designed a convolutional neural network (CNN) classifier to build the personalized mapping from static ultrasound images to eight different hand gestures. To evaluate the performance of the proposed CNN classifier, an ultrasound images dataset and labeled gestures from four young healthy participants were collected and analyzed. Results from offline intra-subject (personalization) validation, quasi-real-time validation, and real-time validation showed high classification accuracy of 99.65%, 97.47%, and 90.83%, respectively. In addition, real-time hand gesture recognition could be executed within 50 ms per image frame by using the proposed CNN classifier. Our findings demonstrated promising real-time hand gesture recognition with high accuracy by using B-mode ultrasound images and the proposed CNN classifier for prosthetic hand control. Copyright (C) 2024 The Authors. This is an open access article under the CC BY-NC-ND license (https://***/licenses/by-nc-nd/4.0/)

关键词： Biomedical and medical image processing and systems Rehabilitation engineering and healthcare delivery Bio-signals analysis and interpretation Man-machine interfaces Bioinformatics Estimation and control in biological systems

来源：评论

学校读者我要写书评

暂无评论

Dense Object Detection in Complex Backgrounds Based on deep learning 30th

Dense Object Detection in Complex Backgrounds Based on Deep ...

引用

30th International Conference on Computational and Experimental Engineering and Sciences

作者： Wan, Xiyuan Luo, Qingdong Zheng, Pengfei Lou, Jingjing Jin, Chaoqun Yiwu Ind Commercial Coll Yiwu Peoples R China Taizhou Vocat & Tech Coll Taizhou Peoples R China East China Univ Sci & Technol Shanghai Peoples R China

ISBN: (纸本)9783031816758;9783031816734;9783031816727

To achieve automatic dense target localization and detection, the following issues need to be addressed: firstly, due to the complex and constantly changing environmental information (Liu Y, Complex background ground based on artificial intelligence Research on multi-objective extraction technology. Xi'an Technological University, Xi'an) collected by mobile cameras and high-altitude drones, pre-processing methods such as filtering, Laplacian enhancement (Yang L, Hai NW, Yu QF, Chin J Aeronaut 36:353-365, 2023) and threshold segmentation are required on the original images to achieve accurate target extraction. Secondly, the shape, size, and color of targets in the image change over time and under changing lighting conditions. Achieving accurate recognition of targets under these conditions is a challenging problem. The team plans to use methods such as dense target tilt correction, overlapping segmentation, and target deduplication. Our team ultimately completed the classification and feature extraction of motion intensive targets under the camera, and based on deep learning, studied the recognition and perception of target behavior in complex environments relying on visual information, ultimately achieving more reliable and universal accurate detection of multiple targets in complex environments.

关键词： deep learning Dense object Complex background Object detection

来源：评论

学校读者我要写书评

暂无评论

Intelligent segmentation and quantification of tunnel lining cracks via computer vision

引用

STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL 2025年第3期24卷 1896-1926页

作者： Feng, Yong Zhang, Xiao-Lei Feng, Shi-Jin Zhang, Wei Hu, Kan Da, Yue-Wu Tongji Univ Urban Mobil Inst Shanghai Peoples R China Tongji Univ Dept Geotech Engn 1239 Siping Rd Shanghai 200092 Peoples R China Minist Educ Key Lab Geotech & Underground Engn Shanghai Peoples R China State Key Lab Disaster Reduct Civil Engn Shanghai Peoples R China Wuxi Water Grp Co Ltd Wuxi Jiangsu Peoples R China

Aiming to automatically, precisely, and rapidly detect tunnel lining cracks from images and extract geometric information for structural condition assessment, this study proposes a novel tunnel lining crack segmentation network (TCSegNet) and establishes a framework for calculating key geometric parameters of cracks. A tunnel lining crack segmentation dataset is first built by conducting on-site inspections of metro tunnels and collecting open-sourced tunnel images. Afterward, the TCSegNet, conforming to the encoder-decoder architectural paradigm, is designed to separate cracks from lining images pixel-to-pixel. An improved ConvNeXt and developed efficient atrous spatial pyramid pooling module constitute the encoder. The skip connections, upsampling modules, and tailored segmentation head form the decoder. Upon the segmentation results of TCSegNet, a computing framework integrating multiple digital image processing techniques is proposed to obtain the length, average width, and maximum width of cracks. The experimental results show that the TCSegNet achieves leading results among several dominant models, with 70.78% mean intersection over union (mIoU) and 57.43% F1 score. Furthermore, the TCSegNet has 32.01 million parameters, requires 55.13 billion floating point operations, and gets 107.28 frames per second, proving that it has low time and space complexities and implements real-time segmentation. Also, the rationality and effectiveness of TCSegNet in alleviating the crack disjoint problem and preserving crack edge details are verified through comparative experiments. In addition, the TCSegNet achieves 71.99%, 70.45%, and 70.23% mIoU in high-resolution image segmentation, robustness, and generalization tests, respectively, demonstrating that it is competent for detecting high-resolution lining images, has a solid resistance to illumination variations, and can be well generalized to other tunnel lining image datasets. Finally, the applicability of the crack quantif

关键词： Tunnel engineering lining cracks deep learning semantic segmentation crack quantification

来源：评论

学校读者我要写书评

暂无评论

Multi-resolution analysis and deep neural network architecture based hybrid feature extraction technique for plant disease identification and severity estimation

引用

EVOLUTIONARY INTELLIGENCE 2024年第2期17卷 1163-1183页

作者： Kirti, K. Rajpal, Navin Yadav, Jyotsna Mondal, Kalyan Kumar Univ Sch Informat Commun & Technol Guru Gobind Singh Indraprastha Univ Delhi India ICAR Indian Agr Res Inst Plant Bacteriol Lab Delhi India

The crop protection techniques using efficient plant disease identification are highly beneficial for Smart Agriculture Management which further decreases the gap between the demand/supply of consumers & farmers. However, an efficient feature extraction technique is desirable to achieve high accuracy, least graphical processing, and low processing time for coloured images. Therefore, this work proposes a hybrid novel feature extraction technique for coloured plant leaf images to overcome these challenges. This work is divided into two parts: 1). Plant disease identification and 2). Plant disease severity estimation. In this technique, multiresolution analysis using large-scale coefficients of Discrete Wavelet Transform (DWT) is performed using orthogonal wavelets. The orthogonal wavelets such as Daubechies, Biorthogonal, Symlets and Coiflets with appropriate vanishing moments are utilized to select significant features. Then, the Eigen decomposition of large-scale DWT-based coloured image features is utilized using Principal Component Analysis (PCA), where only the Eigenvalues with large variance are considered for further feature selection. Several experiments are carried out for the selection of felicitous sub-bands & decomposition levels. The deep Neural network (DNNs) architectures such as AlexNet, ResNet 101 and Inception V3 are incorporated to classify the obtained features. The DNNs operate on Convolutional Neural Networks, which utilize the weights from pre-learned imageNet Database classes to classify various classes. The combination of DNNs with Transfer learning and Fine Tuning is made to tune the parameters for better efficacy. Experiments were performed on the PlantVillage Database, which contains healthy/diseased images of plants such as Apple, Corn, Grapes and tomato leaves. A new database is created by capturing the images of Rice Crop Leaves, including healthy and diseased leaves. A total of 8000 images are processed for disease identification. The

关键词： deep neural networks Discrete wavelet transform Plant disease detection Precision agriculture Multiresolution analysis Smart farming

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：