检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Alehdaghi, Mahdi Shamsolmoali, Pourya Cruz, Rafael M.O. Granger, Eric LIVIA ILLS Dept. of Systems Engineering ETS Montreal Canada Dept. of Computer Science University of York United Kingdom

A key challenge in visible-infrared person re-identification (V-I ReID) is training a backbone model capable of effectively addressing the significant discrepancies across modalities. State-of-the-art methods that generate a single intermediate bridging domain are often less effective, as this generated domain may not adequately capture sufficient common discriminant information. This paper introduces Bidirectional Multi-step Domain Generalization (BMDG), a novel approach for unifying feature representations across diverse modalities. BMDG creates multiple virtual intermediate domains by learning and aligning body part features extracted from both I and V modalities. In particular, our method aims to minimize the cross-modal gap in two steps. First, BMDG aligns modalities in the feature space by learning shared and modality-invariant body part prototypes from V and I images. Then, it generalizes the feature representation by applying bidirectional multi-step learning, which progressively refines feature representations in each step and incorporates more prototypes from both modalities. Based on these prototypes, multiple bridging steps enhance the feature representation. Experiments1 conducted on V-I ReID datasets indicate that our BMDG approach can outperform state-of-the-art part-based and intermediate generation methods, and can be integrated into other part-based methods to enhance their V-I ReID performance. © 2024, CC BY.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Novel Approach for Detecting Brain Neoplasm Using Deep Learning Technique 3

Novel Approach for Detecting Brain Neoplasm Using Deep Learn...

引用

3rd IEEE International Conference on Intelligent Techniques in Control, Optimization and Signal Processing, INCOS 2024

作者： Sridevi, S. Hemalatha, K. Mariammal, G. Vijayalakshmi, V. Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Dept. of Computer Science and Engineering Chennai India Bannari Amman Institute of Technology Sathyamangalam Dept. of Bio Medical Engineering Erode India

ISBN: (纸本)9798350361186

One of the most grinding and significant provoca- tions in the domain of handling images related to medicine is the segmentation of brain neoplasms due to physical categorization with the assist of humans might provoke inappropriate prognosis also an evaluation. Also whenever you've got an abundance of information required being processed, it is a challenging job. sorted through. The extraction of neoplasm regions from pictures becomes difficult since brain neoplasms have a wide variety of appearances and resemble normal tissues. proposed a technique for extracting brain neoplasms from 2D MRIs of the brain using the CNN clustering algorithm, which was then used to train conventional classifiers and a convolutional neural network. For the purpose of this work, a live collection with diverse Neoplasm dimensions, places shapes, and intensity levels of images was used. Sci-kit-Learn was used to implement Support Vector Machine (SVM) and Random Forest, two traditional classifiers, in the traditional classifier section. Then, due to it performing better than the conventional ones, went we now turn to Convolutional Neural Networks (CNN), that are constructed by using Tensorflow along with Keras. The accuracy rate achieved by CNN was 88.0%, which is impressive. © 2024 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Latent Semantic Analysis based Real-world Application of Topic Modeling: A Review Study 2

Latent Semantic Analysis based Real-world Application of Top...

引用

2nd International Conference on Artificial Intelligence and Smart Energy, ICAIS 2022

作者： Gupta, Isha Chatterjee, Indranath Gupta, Neha Mriirs Faculty of Computer Applications Faridabad India Tongmyong University Dept. of Computer Engineering Busan Korea Republic of

ISBN: (纸本)9781665400527

Today, we are living in the digital world. Much information is available on the internet about| every topic. But merely the availability of information is not enough. There is a need for automated tools that can extract hidden patterns or topics from a large number of files or data. Topic modeling is one tool available in data mining that plays a significant role in finding those unknown topics. The paper's objective is to study the importance of topic modeling and discuss the various works done in this field. The author concludes the various fields where more work is required. © 2022 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Technique for Evaluating the Effectiveness of the Information Security Department 16

Technique for Evaluating the Effectiveness of the Informatio...

引用

16th IEEE International Scientific and Technical Conference Actual Problems of Electronic Instrument engineering, APEIE 2023

作者： Belov, Viktor M. Pestunova, Tamara M. Belova, Nadezhda V. Kosov, Dmitry L. Information Security Dept. Novosibirsk State Technical University Novosibirsk Russia Computer Systems Dept. Novosibirsk State University Novosibirsk Russia Economics and Planning Dept. PJSC Rosneft Oil Company Moscow Russia

ISBN: (纸本)9798350330885

This article proposes a technique that establishes the procedure for evaluating the level of efficiency of the information security department (an employee performing information security functions). The technique uses performance evaluation criteria based on the apparatus of fuzzy logic, the composition of fuzzy relations. The technique describes the procedure for evaluating the effectiveness of the information security department (information security officer) during audits in the area of "Organization and state of work on information protection", self-assessment of the effectiveness of work. The method of assessing the level of efficiency consists in presenting with the help of a set of measurements (both at the quantitative and qualitative level) the features collected to build a classification of the effectiveness of the information security department (information security officer). Based on a set of measurements of signs, the decision-maker must determine (classify) the effectiveness of work using the criteria for assessing the quality of their work. In the future, the methodology can be expanded for additional purposes of predicting the level of security of informatization objects. © 2023 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Virtual Reality Based Medical Training Simulator and Robotic Operation System

Virtual Reality Based Medical Training Simulator and Robotic...

引用

2022 International Conference on Recent Progresses in Science, engineering and Technology, ICRPSET 2022

作者： Rahman, Md. Mustafizur Ishmam, Md Fatin Hossain, Md. Tanvir Haque, Md. Emdadul Dept. of Information and Communication Engineering Bangladesh University of Rajshahi Dept. of Computer Science and Engineering Rajshahi6205 Bangladesh

ISBN: (纸本)9798350332704

In the medical sector learning about the human body and practicing surgeries, the medical students need a real human body which is very costly. Repeatedly practice or undone after practice is almost impossible with the human body. Human anatomy, physiology, etc. subjects are very hard to understand and learn from books and images. Learning surgery on living humans is sometimes unethical and trainee surgeons has to struggle to learn rare surgeries. To solve these problems we use virtual reality (VR) to build a medical training simulator that can provide medical students better understanding and learning about the human body through 3D models of the body. Students can get an interactive learning experience and see how organs work from the molecular level and perform several operations. Since this is a simulation everything a new learner will do can be undone and re-practiced. On the other hand skilled surgeons are not available in all places in the country especially in remote places. A patient who needs the critical surgical treatment has to travel to a city to take the treatment by the expert surgeons. To solve this problem in this project we also connect a robot to our VR based system. A skilled surgeon can operate in the virtual environment from a distant place using the robot. The vision system of the robot can be realized by the surgeon through VR headset and he or she feels like, he or she is present there. Surgeons can control the robotic hands and tools necessary to perform the surgery from the remote place. This will reduce the trouble of patients traveling from remote places to city area for surgery. We implement the demo system in the lab and show some results which indicate the applicability of the proposed system. © 2022 IEEE.

关键词： Robotic surgery

来源：评论

学校读者我要写书评

暂无评论

Experimental Demonstration of Hybrid LED-LD Based UOWC Link 2

Experimental Demonstration of Hybrid LED-LD Based UOWC Link

引用

2nd International Conference on Microwave, Optical and Communication engineering, ICMOCE 2023

作者： Krishna, Gopal Mathew, Jimson Santhanakrishnan, T. Gupta, Sumanta Indian Institute of Technology Patna Dept. of Electrical Engineering India Indian Institute of Technology Patna Dept. of Computer Science and Engineering India DRDO Naval Physical and Oceanographic Laboratory India

ISBN: (纸本)9798350321609

In this paper, we have designed and experimentally characterized a hybrid light emitting diode (LED) and laser diode (LD) based underwater optical wireless communication (UOWC) link, which can be suitable for providing high-speed optical connectivity between onshore and submerged systems. An LED-based system is efficient for short-range communication whereas an LD-based system is for long distances. In this paper, we have experimentally demonstrated the establishment of a hybrid LED-LD UOWC link at a data rate of 10 Mbps over a 2.5 m water channel using a return to zero on-off keying (RZ-OOK) modulation scheme using an FPGA-based transmitter and receiver. We have also used offline data processing to quantify the bit error rate (BER) performance as a function of received optical power for tap water and turbid water channels. For the tap water channel, we have achieved a BER as low as 3.5 ×10-6 at - 35.1 dBm received optical power, whereas for turbid water the achieved BER is 1.8 × 10-5 at -31. 6 dBm of received optical power in absence of air bubble. In turbid water (and tap water) channel in presence of air bubbles the minimum achievable BER is around 3 × 10-4 (3 × 10-4) for -31.5 dBm (-35.5 dBm) of received power. © 2023 IEEE.

关键词： Field programmable gate arrays (FPGA)

来源：评论

学校读者我要写书评

暂无评论

TUR2SQL: A Cross-Domain Turkish Dataset For Text-to-SQL 8

TUR2SQL: A Cross-Domain Turkish Dataset For Text-to-SQL

引用

8th International Conference on computer Science and engineering, UBMK 2023

作者： Kanburoglu, Ali Bugra Tek, F. Boray Işik University Huawei R&d Center Dept. of Computer Engineering İstanbul Turkey Istanbul Technical University Dept. of Artificial Intelligence and Data Engineering İstanbul Turkey

ISBN: (纸本)9798350340815

The field of converting natural language into corresponding SQL queries using deep learning techniques has attracted significant attention in recent years. While existing Text-to-SQL datasets primarily focus on English and other languages such as Chinese, there is a lack of resources for the Turkish language. In this study, we introduce the first publicly available cross-domain Turkish Text-to-SQL dataset, named TUR2SQL. This dataset consists of 10,809 pairs of natural language statements and their corresponding SQL queries. We conducted experiments using SQLNet and ChatGPT on the TUR2SQL dataset. The experimental results show that SQLNet has limited performance and ChatGPT has superior performance on the dataset. We believe that TUR2SQL provides a foundation for further exploration and advancements in Turkish language-based Text-to-SQL research. © 2023 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

ML based Pre-Diagnosis Framework for early detection of Neurodevelopmental Disorders 4

ML based Pre-Diagnosis Framework for early detection of Neur...

引用

4th International Conference on Advancement in Electronics and Communication engineering, AECE 2024

作者： Pandey, Sumit Sharma, Srishti Agarwal, Anant The Northcap University School of Engineering and Technology Dept. of Computer Science Haryana Gurugram122017 India Aligarh Muslim University Zakir Husain College of Engg. and Tech Dept. of Electronics Engineering Uttar Pradesh Aligarh202001 India

ISBN: (纸本)9798350364729

Neurodevelopmental disorders cover a wide range of behavioral and cognitive symptoms mapped to various conditions. Early diagnosis and planned interventions are crucial for improving outcomes and the quality of life for people suffering from such disorders. This research work talks about a detailed framework with the objective of assisting in the early pre-diagnostic phase of neurodevelopmental disorders. The motivation for the development of a ML-based framework in pre-diagnosis of neurodevelopmental disorders is the effective and zero-cost getting diagnosed. This framework will integrate knowledge from diverse domains to systematize and make recognition of early signs and symptoms associated with neurodevelopmental disorders much more efficient. Additionally, the framework emphasizes two very critical parameters: accessibility and cost-effectiveness. The developed pre-diagnostic mechanism can be implemented in a diversified variety of scenarios, especially with limited resources, towards zero cost. The implementation of advanced machine learning techniques can assist in the analysis of the response to a customized questionnaire in identifying patterns that may suggest a risk of neurodevelopmental disorders. This proactive approach not only supports and facilitates early detection and intervention but also seeks to add towards the overall well-being of patients with these disorders. © 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

Leg Movement Controlled Robotic Leg for Menace Preclusion

Leg Movement Controlled Robotic Leg for Menace Preclusion

引用

2024 IEEE SmartBlock4Africa, SmartBlock4Africa 2024

作者： Sitti, Mensah Antwiwaa, Anita Danquah-Amoah, Archibald Govindan, Kannan Martey, Ezekiel Mensah University of Mines and Technology Dept of Computer Science and Engineering Tarkwa Ghana All Nations University College Dept. of Electronics and Communication Engineering Koforidua Ghana Accra Technical University Dept. of Medical Laboratory Accra Ghana University of Health and Allied Sciences Dept. of Orthotics and Prosthetics Ho Ghana

ISBN: (纸本)9798350376838

Movement of autonomous humanoid is a key area, as we gradually accept robots. This research proposes an 'Artificial Human Leg Model' which can precisely follow the real movements of human leg by means of image processing thereby avoiding sensor noise and damage and providing more adjustable response. Webcams are used to capture the leg images followed by Morphological image processing to extract information from human leg movement. The four degrees of freedom implemented in the proposed method are rotation of base, leg like motions, knee like motions, and steps motions. The robot leg is controlled successfully using human leg movement, where the system can travel a distance, manoeuvre in an uneven surface and move the robot from one place to other with precision. The user can control the leg entirely based on his/her intentions to accomplish the required task. The proposed real-time method of implementing human leg movement also ensures immediate control and performance. Distant control of the robotic leg does not only ensure complete safety to human, but also allows flexibility in controlling distant objects. The overall goal of the proposed system is to guarantee safety, flexibility and accuracy in movements when performing sensitive tasks in day-to-day life activities thus offering suitability for menace preclusion. © 2024 IEEE.

关键词： Motion capture

来源：评论

学校读者我要写书评

暂无评论

AI-Driven Panic Detection and Alert System Using Smartwatch and LLM Model

AI-Driven Panic Detection and Alert System Using Smartwatch ...

引用

International Conference on Parallel Distributed and Grid Computing (PDGC)

作者： Balaram Pal Anubhab Gupta Soumyajit Paul Munshi Mustafijur Rahaman Dept. of Computer Science Engineering Bengal Institute of Technology Kolkata India Dept. of Computer Science Engineering GCECT Kolkata India

ISBN: (数字)9798331521349

ISBN: (纸本)9798331521356

Technologically, most existing systems cannot detect distress autonomously through pattern recognition, further compounding the challenge of providing timely help. This research paper identifies the significant gap in the market for an automated, real-time alert system that provides a solution that does not require manual engagement. In response, an LLM-based Smartwatch-Based Panic Alert System (SBPAS) is proposed to integrate into smartwatches that continuously monitor heart rate, motion, and location biometric signals. By applying advanced machine learning models, the system detects abnormal patterns consistent with panic or distress and automatically triggers alerts to emergency contacts and local authorities. This eliminates the need for manual intervention, offering a faster, more reliable response in critical situations. The significance of this solution lies in its ability to reduce response times, improve the likelihood of timely assistance, eliminate the chance of not being notified about an emergency, and close the existing technological gap in safety infrastructure, particularly in regions where crimes against individuals are prevalent.

关键词： Biometrics Heart rate Manuals Machine learning Real-time systems Time factors Reliability Biomedical monitoring Wearable devices Monitoring

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：