检索结果-内蒙古大学图书馆

Wrist joint synovial hypertrophy and effusion detection in musculoskeletal ultrasound images using self-attention U-Net

引用

Multimedia Tools and Applications 2024年第41期83卷 89317-89334页

作者： Chang, Chuan-Wang Chang, Chuan-Yu Zhu, Yu-Xian Wang, Sz-Tsan Department of Computer Science and Information Engineering National Chin-Yi University of Technology Taichung Taiwan Department of Computer Science and Information Engineering National Yunlin University of Science and Technology Douliu Taiwan Division of Allergy Immunology and Rheumatology Dalin Tzu Chi Hospital Buddhist Tzu Chi Medical Foundation Dalin Chiayi Taiwan

Skeletal muscle ultrasound has emerged as a pivotal imaging modality in rheumatology clinics, offering unparalleled advantages such as radiation-free imaging, safety, and dynamic examination capabilities. However, its reliance on operator expertise often leads to inconsistent interpretations and diagnostic variability. In this study, we present a novel diagnostic system aimed at detecting rheumatoid arthritis (RA) in the wrist joint, with a focus on identifying synovial hypertrophy and effusion using musculoskeletal ultrasound images. Leveraging deep learning techniques, specifically semantic segmentation models, we introduce SEAT-UNet, which combines the U-Net architecture with a self-attention mechanism to enhance the accuracy of lesion classification and localization. SEAT-UNet addresses the challenge of discontinuous dispersion encountered in conventional segmentation models, particularly when delineating lesion areas. Our experimental results demonstrate exceptional performance, achieving a sensitivity and Dice coefficient of 100% and 84%, respectively, in synovial hypertrophy detection, and 86% sensitivity with an 84% Dice coefficient in effusion detection. These findings underscore the potential of SEAT-UNet as a valuable tool for early RA diagnosis, offering improved patient outcomes and facilitating more effective disease management strategies. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Real-Time Gesture Based Sign Language Recognition System

Real-Time Gesture Based Sign Language Recognition System

引用

2024 International Conference on Advances in Data engineering and Intelligent Computing Systems, ADICS 2024

作者： Debnath, Jeet Praveen Joe, I.R. School Of Computer Science And Engineering Vellore Institute Of Technology Chennai India

ISBN: (纸本)9798350364828

Sign language is a vital mode of communication for the deaf and dumb community. This research presents a robust and real-time Gesture-Based Sign Language Detection System that leverages computer vision and deep learning techniques. The system is designed to recognize and interpret American Sign Language (ASL) gestures, enabling efficient communication between individuals who are proficient in ASL and those who are not. The core of the system utilizes Python, OpenCV (Open-Source computer Vision Library), and MediaPipe Holistic for real-time hand and body pose estimation. By accurately tracking the movements of hands and key body parts, the system captures the nuances of sign language gestures. The captured data is then fed into a Long Short-Term Memory (LSTM) neural network, which excels in sequence modeling tasks. The LSTM model is trained on a comprehensive dataset of ASL gestures, encompassing a wide range of signs and expressions. Transfer learning techniques are also employed to fine-tune the model for improved performance in sign language recognition. The model's architecture allows it to learn the temporal dependencies and context inherent in sign language, making it capable of recognizing gestures within sentences or phrases. The system's evaluation demonstrates its effectiveness in real-world scenarios, achieving high accuracy and low latency in sign language recognition. It opens new avenues for accessible and inclusive communication, aiding both deaf and hearing individuals in bridging the communication gap. Future work may explore the integration of natural language processing (NLP) to facilitate two-way communication between sign and spoken language. In conclusion, this Gesture-Based Sign Language Detection System represents a significant step towards harnessing the power of computer vision and deep learning to make sign language more accessible and inclusive in various domains, including education, accessibility, and social interaction. © 2024 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Examination of Toxic Tweet Classification on Twitter 9

A Comprehensive Examination of Toxic Tweet Classification on...

引用

9th International Conference on science, technology, engineering and Mathematics, ICONSTEM 2024

作者： Gokulnath, Shreejith Suthraye Jayaraman, Shrikar Nathezhtha, T. School of Computer Science And Engineering Vellore Institute of Technology Chennai India

ISBN: (纸本)9798350365092

In recent times, daily interactions have been taking place largely on online platforms, with forms of interactions such as ***,an important problem lies in the understanding of these messages. These textual interchanges can be interpreted in multiple different ways by different people which can cause significant misunderstandings and negative perceptions. The aim of this project is to address this challenge and overcome it by developing a system that is capable of identifying the negativity and toxicity in the tone of the messages. Through rigorous analysis the model aims to identify parts or sequences of a message that might appear as offensive or negative, giving the user constructive feedback on how they can better improve their messages. This helps users to write messages that are not only clearer but also easily received in the intended manner by their audience, thereby increasing the effectiveness of the communication process and reducing misunderstandings. The main goal of this paper is to contribute to the development of an inclusive and understanding online community by using techniques of Natural Language Processing and Machine Learning to solve the complexities of human communication. © 2024 IEEE.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Quantum-Based QKD and Sugar-Salt Encryption Approach in Cloud-Fog Computing to Strengthen Protection of Online Banking Data

Quantum-Based QKD and Sugar-Salt Encryption Approach in Clou...

引用

2024 IEEE International Conference on Information technology, Electronics and Intelligent Communication Systems, ICITEICS 2024

作者： Turjya, Sapthak Mohajon Singh, Raj Sarkar, Pritam Swain, Sujata Bandyopadhyay, Anjan Kalinga Institute of Industrial Technology Computer Science and Engineering Bhubaneswar India

ISBN: (纸本)9798350382693

Cloud computing is swiftly developing, delivering a diverse spectrum of computing utilities like databases, storage, and servers. It transfigures the way data storage is done and securely circulated among numerous users. Nevertheless, enabling privacy and restricting identity breaches when availing cloud utilities is essential. Fog Computing, an updated version of the cloud framework, utilizes a decentralized architecture to maintain data between the cloud and the data sources. This procedure seeks to improve performance, network connectivity, and most importantly provide security. Quantum key distribution (QKD) is employed in fog computing architectures to develop and circulate keys. The employed keys change their quantum configurations, enabling them to be highly safe. Further, an intermediary maintaining the key can identify any endeavor by illegal entities to operate it. In this research article, we utilize sugar-salt encryption, an approach designed to oppose brute-force invasions efficiently. This encryption procedure delivers erroneous data in reply to wrong password/key insertions, improving data protection, which is pivotal in online based banking data protection. © 2024 IEEE.

关键词： Fog computing

来源：评论

学校读者我要写书评

暂无评论

Prediction of Common Diseases Lurking in Smokers and Alcoholics and its Impact on their Health

Prediction of Common Diseases Lurking in Smokers and Alcohol...

引用

2024 International Conference on Integration of Emerging Technologies for the Digital World, ICIETDW 2024

作者： Jeevan, B. Koushik Vishnu, D. Jagadish, R. Praveen Joe, R.I. Vellore Institute of Technology School of Computer Science and Engineering Chennai India

ISBN: (纸本)9798331541170

Smoking and alcohol intake have become very common in modern society, leading to many complications and health problems, including respiratory illnesses and cardiovascular diseases. Early intervention and targeted healthcare interventions are very much needed to mitigate these risks and improve patient outcomes. This research addresses the need by using patient data to predict potential future health risks. By analysing 19 health and lifestyle factors, including smoking, alcohol consumption, diabetes, hypertension, haemoglobin and such factors, the study developed a predictive model to identify diseases that patients may currently have or develop in the near future. After extensive hyperparameter tuning and analysis, the study utilized a deep learning model optimized with the ReduceLROnPlateau technique, Adam as the optimizer and binary cross-entropy as the loss function. With the best parameters, including a learning rate of 0.001, 200 epochs, and a batch size of 64, the model achieved the highest accuracy of 94.69%. Additionally, to assess and compare the efficiency of the deep learning model, the study explored Random Forest algorithm due to its robustness in handling complex datasets. Initially focused on smokers and alcoholics, the study expanded to include non-smokers and non-alcohol consumers, demonstrating the model's broad applicability. These findings highlight the model's strong predictive capability and its potential for widespread use in preventive healthcare. Further, the current study uses dashboard visualization that reveals common disorders that affect smokers and drinkers, providing the valuable insights on targeted treatments to improve patient care and outcomes in population. © 2024 IEEE.

关键词： Network security

来源：评论

学校读者我要写书评

暂无评论

AI-Enhanced Image Translation for Seamless Communication in Global Tourism

AI-Enhanced Image Translation for Seamless Communication in ...

引用

2024 International Conference on Advances in Data engineering and Intelligent Computing Systems, ADICS 2024

作者： Rubasri, S. Shafiur Rahman, N.F. Syed Khalid, J. Vignesh Balaji, R. Prasath, R. Krishna, B.V. Kcg College Of Technology Department Of Computer Science & Engineering Chennai India

ISBN: (纸本)9798350364828

An inventive program called Jargon Lens was created to meet the various language requirements of international visitors. AI-enhanced image translation technology, jargon lens empowers travelers to effortlessly overcome language barriers by seamlessly translating text in images to their preferred language. The core of this project is to create a user-friendly mobile application that can recognize and translate text within images accurately and in real-time. This will enable users to easily understand and communicate in different languages by simply using the Smartphone camera, users can scan or upload and translate text in real-time, ensuring a swift and accurate understanding of their surroundings. Beyond personal advantages, Jargon Lens supports constructive encounters that promote understanding and cultural interchange between visitors and local populations. As a technological solution, Jargon Lens aligns with the evolving landscape of global tourism, where diversity and inclusivity are paramount. This technology combine's computer vision such as optical character recognition (OCR). OCR algorithms analyze the image to recognize and extract the textual content and natural language processing (NLP) to perform the translation. In conclusion, Jargon Lens emerges as a transformative tool for tourists, offering a sophisticated yet easy-to-use solution for navigating language barriers. With its focus on real-time multilingual image translation, Jargon Lens stands as a key enabler for enriched and hassle-free travel experiences, bridging communication gaps in a rapidly evolving global tourism ecosystem. © 2024 IEEE.

关键词： Optical character recognition

来源：评论

学校读者我要写书评

暂无评论

A New Hybrid Optimization Algorithm for Reliable Clustering Routing in Wireless Sensor Network 2

A New Hybrid Optimization Algorithm for Reliable Clustering ...

引用

2nd IEEE International Conference on Integrated Intelligence and Communication Systems, ICIICS 2024

作者： Joshi, Vijayalaxmi Biradar, Mallanagouda Deepak, B.N. Biradar, Manjula Department of Computer Science and Engineering SJB Institute of Technology Bangalore India Department of Computer Science and Engineering Sharnbasva University Kalaburagi India Department of Information Science and Engineering Sharnbasva University Kalaburagi India

ISBN: (纸本)9798331504960

A wide-spread network of tiny sensor nodes is often what makes up a wireless sensor network (WSN). One or more sinks sometimes referred to as base stations (BSs), collect the data from these SNs. Compared to a standard ad-hoc network, WSNs are significantly more complex due to their intrinsic routing feature. In a wireless sensor network (WSN), nodes perform diverse tasks and are typically organized into multiple clusters according to specific requirements or metrics. This paper develops a new cluster head selection (CHS) technique for WSN using a novel hybrid algorithm called Secretary Bird Integrated Walrus Optimization (SBI-WaO). By considering QoS, throughput, network quality (reliability), energy, and risk, the best possible CHS is carried out. The final findings demonstrate the SBI-WaO algorithm's effectiveness in terms of energy, throughput, etc. With the recommended SBI-WaO approach, a high residual energy of 10.3938 is obtained. © 2024 IEEE.

关键词： Sensor nodes

来源：评论

学校读者我要写书评

暂无评论

Transferable Adversarial Facial Images for Privacy Protection 24

Transferable Adversarial Facial Images for Privacy Protectio...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Li, Minghui Wang, Jiangxiong Zhang, Hao Zhou, Ziqi Hu, Shengshan Pei, Xiaobing Software Engineering Huazhong University of Science and Technology Wuhan China Computer Science and Technology Huazhong University of Science and Technology Wuhan China Cyber Science and Engineering Huazhong University of Science and Technology Wuhan China

ISBN: (纸本)9798400706868

The success of deep face recognition (FR) systems has raised serious privacy concerns due to their ability to enable unauthorized tracking of users in the digital world. Previous studies proposed introducing imperceptible adversarial noises into face images to deceive those face recognition models, thus achieving the goal of enhancing facial privacy protection. Nevertheless, they heavily rely on user-chosen references to guide the generation of adversarial noises, and cannot simultaneously construct natural and highly transferable adversarial face images in black-box scenarios. In light of this, we present a novel face privacy protection scheme with improved transferability while maintain high visual quality. We propose shaping the entire face space directly instead of exploiting one kind of facial characteristic like makeup information to integrate adversarial noises. To achieve this goal, we first exploit global adversarial latent search to traverse the latent space of the generative model, thereby creating natural adversarial face images with high transferability. We then introduce a key landmark regularization module to preserve the visual identity information. Finally, we investigate the impacts of various kinds of latent spaces and find that F latent space benefits the trade-off between visual naturalness and adversarial transferability. Extensive experiments over two datasets demonstrate that our approach significantly enhances attack transferability while maintaining high visual quality, outperforming state-of-the-art methods by an average 25% improvement in deep FR models and 10% improvement on commercial FR APIs. © 2024 ACM.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Improving Multi-Domain Task-Oriented Dialogue System with Offline Reinforcement Learning

Improving Multi-Domain Task-Oriented Dialogue System with Of...

引用

2024 IEEE International Conference on Big Data, BigData 2024

作者： Prajapat, Dharmendra Toshniwal, Durga Indian Institute of Technology Dept of Computer Science and Engineering Roorkee India

ISBN: (纸本)9798350362480

Task-oriented dialogue (TOD) system is designed to accomplish user-defined tasks through dialogues. The TOD system has progressed towards end-to-end modeling by leveraging pre-trained large language models. Fine-tuning the pre-trained language models using only supervised learning leads to the exposure bias and token loss problem and it deviates the models from completing the user's task. To address these issues, we propose a TOD system that leverages a unified pre-trained language model, GPT-2, as a base model. It is optimized using supervised learning and offline reinforcement learning (RL). The issues in the TOD system are mitigated using a non-differentiable reward function. The reward is calculated using the weighted sum of the success rate and BLEU evaluation metrics. The success rate and BLEU metrics in reward calculation guide the language model for user task completion while ensuring a coherent and fluent response. Our model is acquired by fine-tuning a pre-trained model on the dialogue-session level which comprises user utterance, belief state, system act, and system response. Experimental results on MultiWOZ2.1 demonstrate that our model increases the inform rate by 1.60% and the success rate by 3.17% compared to the baseline. © 2024 IEEE.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

Enhanced Crop Yield Prediction by Machine Learning Techniques 3

Enhanced Crop Yield Prediction by Machine Learning Technique...

引用

3rd IEEE International Conference of Electron Devices Society Kolkata Chapter, EDKCON 2024

作者： Das, Papiya Pal, Abhranil Mullick, Rakteem Singha, Sayak Nath, Subhrapratim Meghnad Saha Institute Of Technology Computer science & Engineering Kolkata India

ISBN: (纸本)9798350374643

India's economy is based primarily on agriculture. The need for agriculture has increased along with India's population growth during the previous ten years. However, climatic changes and poor irrigation is a major threat to agricultural practices in India. Many farmers practice agriculture solely based on their experience which in a way reduces productivity and effectiveness. At present, machine learning techniques provide effective recommendations and predictions, helping to scale the production of crops across the country effectively. There are many machine learning based algorithms and techniques which help us accomplish effective solutions to such problems. In this work, Random Forest algorithm proposed which predicts the crops based on weather factors and soil factors. The weather factors include Rainfall, Temperature and Humidity. The soil factors include the chemical composition of the soil such as NPK values and pH of the soil. Real-time data used for the weather conditions, which makes the model robust and scalable. The test results will show two crop names along with its yield which are best suited for the respective conditions. Random Forest Classifier used for crop name recommendation and Random Forest Regressor for crop yield prediction. These results will assist farmers to grow crops effectively and also increase production of crops. © 2024 IEEE.

关键词： Adversarial machine 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：