检索结果-内蒙古大学图书馆

An In-Depth Study of Handwritten Digit Recognition Methods: An Exploration of Deep Learning and Machine Learning Approaches

An In-Depth Study of Handwritten Digit Recognition Methods: ...

引用

2023 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems, ICSES 2023

作者： Singh, Nongmeikapam Thoiba Tiwary, Priyanshu Chanu, Asem Debala Chandigarh University Department of Computer Science and Engineering Punjab India Indian Institute of Information Technology Department of Computer Science and Engineering Assam Guwahati India

ISBN: (纸本)9798350319200

This article explains how deep learning and machine learning techniques are being utilized to replace humans in a variety of activities, like item classification and sound addition. It emphasizes handwritten text recognition (HTR), which involves computers interpreting handwritten content from sources like paper documents and touchscreens. For recognizing handwritten digits, the study analyzes three algorithms: support vector machines (SVM), multilayer perception (MLP), and convolutional neural networks (CNN). This comparison aims to evaluate the correctness and execution times of different approaches. The evaluation is based on parameters like the dataset used, the number of epochs, algorithm complexity, accuracy, and hardware specifications. The study compares the effectiveness of CNN, Multilayer Perception (MLP), and Support Vector Machines using handwritten digits from the MNIST dataset. It evaluates these algorithms based on criteria such as runtime, complexity, rate of precision, the count of epochs, and the quantity of hidden layers specifically within deep learning techniques. SVM had the highest training data accuracy, but CNN outperformed others on the test dataset. The article's primary goal is to identify the most precise model for recognizing handwritten digits. It achieves this by assessing the performance of different algorithms and comparing them based on various characteristics. © 2023 IEEE.

关键词： convolutional neural networks execution time handwritten digits multilayer perceptron overfitting visualization

来源：评论

学校读者我要写书评

暂无评论

Lasso Regression Model of HSV Color Space Masking Vision Based Whole Coffee Bean Processing Agtron Measurement

引用

Journal of institute of Control, Robotics and Systems 2024年第11期30卷 1245-1253页

作者： Lee, Ye Chan Noh, Yi Hyun Noh, Yun Seo Kim, Ju Song Ryu, Kwangyeol Kim, Byeong Hak Department of Industrial Engineering Pusan National University Korea Republic of Department of Computer Science and Engineering Pusan National University Korea Republic of AI Research Center for Manufacturing Systems Korea Institute of Industrial Technology Korea Republic of

The rapid increase in coffee consumption has led to a significant expansion in production scale and variety within the agricultural regions of the global coffee belt. Recent coffee harvested in varies specious and processed in various ways, and the functionality of processing equipment is advanced to optimize its quality. In particular, coffee roasting equipment has evolved to incorporate complex structures and advanced electronic, mechanical, and heat/air control technologies for processing harvested green coffee beans. One of the most critical quality factors in the roasting process is the degree of roast, which is measured by the bean color (Agtron) value. The conventional method for Agtron measurement involves grinding the beans and using vision sensors. However, the conventional method shows difficult and time-consuming. The method for measuring the color of whole beans without grinding is simpler, but it has limitations in accuracy due to variations in shading and background images. This study proposes a method to accurately measure the Agtron color of whole coffee beans without grinding. The proposed method employs a preprocessing technique that separates the measurement area from the background area through color histogram analysis. To enhance measurement accuracy, we suggest using a Lasso regression model, which mitigates overfitting problems. Experimental results demonstrate that the proposed method achieves an improved accuracy compared with traditional regression methods and neural network-based approaches. © ICROS 2024.

关键词： Grinding (machining)

来源：评论

学校读者我要写书评

暂无评论

Secured Storage of Information Using Audio Steganography 7th

Secured Storage of Information Using Audio Steganography

引用

7th International Conference on Emerging Research in Computing, Information, Communication and Applications, ERCICA 2022

作者： Sowmya, M.R. Shreenath, K.N. Shetty, Saritha Shetty, Savitha Wajid, Salman Kantharaj, Yashas Department of Computer Science and Engineering NMIT Yelahanka Bangalore India Department of Computer Science and Engineering Siddaganga Institute of Technology Tumakuru India Department of MCA NMAM Institute of Technology Karkala India Department of CSE NMAM Institute of Technology Karkala India

ISBN: (纸本)9789811954818

The use of smartphones and cloud storage has expanded dramatically as technology progresses. The data that are kept and used are now more vulnerable to breaches. Using the Internet services, the end users’ data files can be accessed and manipulated from any other computer. If the secrecy of the information is of very high value, it should be secured. If we want to prevent unauthorized disclosure or change of information, it must be protected. The old cryptographic methods rely on keys and have a point of failure, and they attract the third party who is on lookout for crucial information. Steganography can be used to tackle this problem. It is a secret method of encrypting data in another form so that it can be securely stored and transferred over any medium. In the following paper, the idea of steganography, implementation methods, and storing the data in an audio file is discussed. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Steganography

来源：评论

学校读者我要写书评

暂无评论

A Robust Object Recognition Using Modified YOLOv5 Neural Network 10

A Robust Object Recognition Using Modified YOLOv5 Neural Net...

引用

10th International Conference on Signal Processing and Integrated Networks, SPIN 2023

作者： Balmik, Archana Barik, Subhasish Nandy, Anup National Institute of Technology Rourkela Department of Computer Science & Engineering Rourkela India

ISBN: (纸本)9781665490993

Human beings can recognize and detect the object in any image or video with ease, but when it comes to machines dealing with the same task, it's quite challenging. In the past years, there has been significant progress in the area of computer vision. However, researchers are still working on enhancing the performance of object recognition algorithms in terms of accuracy and speed. This paper presents one of the greatest CNN representatives, You Only Look Once (YOLO), which tackles object detection in the simplest and most efficient way. This study primarily focuses on object recognition with modified YOLOv5. Here, we have implemented the model on a dataset comprising five objects. We have analyzed various preprocessing and feature extraction techniques to enhance the performance of the proposed model. The highest mean average precision (mAP) achieved by the modified YOLOv5 is 90.7%. The recognition performance of the modified YOLOv5 is compared with the YOLOv4 algorithm. It is also compared with various conventional object detection algorithms outside the YOLO family, such as Support vector machine (SVM), deep Convolutional neural network (CNN), and EfficientNet. The results proved to be promising as compared to the other state-of-the-art techniques. © 2023 IEEE.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Homomorphic Encryption based Federated Learning for Financial Data Security 4

Homomorphic Encryption based Federated Learning for Financia...

引用

4th International Conference on Computing and Communication Systems, I3CS 2023

作者： Dhiman, Shalini Nayak, Sumitra Mahato, Ganesh Kumar Ram, Anil Chakraborty, Swarnendu Kumar National Institute of Technology Department of Computer Science & Engineering Arunachal Pradesh India

ISBN: (纸本)9798350323771

Federated Learning is a distributed machine learning technique that enables on-device training without exchanging the sensitive data over the centralized server. In this paper, Federated Learning is used to train financial data models on-devices with the help of IoT applications in financial or business systems. This creates more advanced and secured financial business models. We have applied the mechanism of homomorphic encryption and cryptographic primitives, including masking and local model protection to prevent any kind of inferring in private financial or business data where multiple attackers usually find a way of model inversion or reconstruction attack. We used datasets of various financial sectors as the primary measure, rather than taking the size of datasets that is generally used in deep learning, to get the correct measurement of the rate of contribution in every session of the local model's training to the training of the global model. If the calculated online number of clients exceeds a predetermined threshold, then the federated learning process will be continued with the dropout-tolerant plan. The security study demonstrates that the suggested solution fulfils data privacy requirements. The costs of computation and communication are also examined theoretically. According to research observations, the proposed approach achieved promising outcomes while assuring privacy preservation when compared to existing schemes. © 2023 IEEE.

关键词： Sensitive data

来源：评论

学校读者我要写书评

暂无评论

Empowering Tourists with Context-Aware Recommendations using GAN 2

Empowering Tourists with Context-Aware Recommendations using...

引用

2nd International Conference on Electronics and Renewable Systems, ICEARS 2023

作者： Stephy, E. Evangelin Rajeswari, M. Karunya Institute of Technology and Sciences Department of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9798350346640

A technology-based solution known as a tourism recommendation system makes suggestions to visitors based on their preferences, prior travels, and experiences. These systems gather data from a variety of sources, including web searches, user reviews, travel history, etc. When there is a lack of data about some users or items, current technologies suffer from sparsity, which can lead to inaccurate recommendations. Another challenge that tourism faces is the diversity issue, which occurs when similarities are valued over preferences. GAN and context-aware recommendation systems are two prominent techniques that are combined in our system. The system's objective is to give personalized recommendations to travelers based on multiple contextual elements. The system creates new recommendations based on the knowledge it has gained by using GANs to identify patterns and connections between various elements of a tourist's setting and their preferences. Synthetic data is generated to supplement the original dataset, which can aid in overcoming the cold-start and sparsity issues. It can also aid in the development of more scalable recommendation systems. © 2023 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

An Improved Structure for Academic Information Services through AI Chatbots

引用

Journal of engineering Science and technology Review 2023年第5期16卷 164-173页

作者： Ula, Mutammimul Hardi, Richki Hipiny, Irwandi Department of Information Systems Faculty of Engineering Universitas Malikussaleh Aceh Indonesia Department of informatics Universitas Mulia Balikpapan Indonesia Faculty of Computer Science and Information Technology Universiti Malaysia Sarawak Sarawak Malaysia

Academic information service is the most critical factor that must be considered in a university. Service quality is an important indicator affecting all academics’ satisfaction and loyalty. Improvement of information services on campus has been a lot of research focusing on academics. Meanwhile, the software aspect has received less attention. Researchers have previously discussed information service frameworks but have not accommodated future needs. Software that must be considered can quickly respond to the user, one of which is a chatbot. Chatbots can be used to solve problems related to question-and-answer services, especially for academic institutions that do not have specific resources to handle this work. With chatbots, redundant user questions can be handled automatically. Chatbots are increasingly being offered as an alternative source of customer service. For a user to use a chatbot for this purpose, the user must trust the chatbot to provide the necessary support. To overcome this knowledge gap, a framework can integrate software, stakeholders, and academics in higher education. Therefore it is necessary to conduct new research to help answer future needs. © 2023 School of Science, IHU. All rights reserved.

关键词： Information services

来源：评论

学校读者我要写书评

暂无评论

The EM Algorithm is Adaptively-Optimal for Unbalanced Symmetric Gaussian Mixtures

引用

Journal of Machine Learning Research 2022年第1期23卷 1-79页

作者： Weinberger, Nir Bresler, Guy The Viterbi Faculty of Electrical and Computer Engineering Technion|Israel Institute of Technology Haifa3200004 Israel Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA02139 United States

This paper studies the problem of estimating the means ± θ2 Rd of a symmetric twocomponent Gaussian mixture δN(θ;I) + (1 - δ) N(-θ;I), where the weights δand 1 - δare unequal. Assuming that δis known, we show that the population version of the EM algorithm globally converges if the initial estimate has non-negative inner product with the mean of the larger weight component. This can be achieved by the trivial initialization θ0 = 0. For the empirical iteration based on n samples, we show that when initialized at θ0 = 0, the EM algorithm adaptively achieves the minimax error rate O ( min n 1 (1-2δ) q d n;1 kθk q d n;- d n )1/4 }) in no more than O ( 1 θ (1-2δ) ) iterations (with high probability). We also consider the EM iteration for estimating the weight δ, assuming a fixed mean θ (which is possibly mismatched to θ). For the empirical iteration of n samples, we show that the minimax error rate ∼O ( 1 θ q d n ) is achieved in no more than O ( 1 θ2 ) iterations. These results robustify and complement recent results of Wu and Zhou (2019) obtained for the equal weights case δ= 1/2. © 2021 Nir Weinberger and Guy Bresler.

关键词： Maximum principle

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive and Systematic Analysis of One-Way Hashing and Its Advancements 3

A Comprehensive and Systematic Analysis of One-Way Hashing a...

引用

3rd International Conference on Intelligent Systems, Advanced Computing, and Communication, ISACC 2025

作者： Chakraborty, Abhilash Biswas, Anupam National Institute of Technology Silchar Department of Computer Science and Engineering Assam Silchar788010 India

ISBN: (纸本)9798331523893

Cryptographic hash functions are fundamental to secure digital interactions. This paper surveys one-way hashing, covering theoretical foundations, historical developments, cryptanalytic challenges, and widely trusted standards like SHA-2, SHA-3, and BLAKE2/BLAKE3. It also explores experimental designs and proprietary functions enhancing password protection, IoT security, and blockchain mechanisms. The study examines tensions between performance and security, classical versus post-quantum resilience, and theoretical rigor versus practical engineering. Hashing is both a mature discipline and an evolving frontier, poised to adapt to changing computational paradigms and sophisticated adversarial methods while meeting growing demands for data integrity and authenticity guarantees. © 2025 IEEE.

关键词： Authentication

来源：评论

学校读者我要写书评

暂无评论

CLAACS-IOD: Certificate-embedded lightweight authentication and access control scheme for Internet of Drones

CLAACS-IOD: Certificate-embedded lightweight authentication ...

引用

作者： Sadhukhan, Dipanwita Ray, Sangram Dasgupta, Mou Rodrigues, Joel J. P. C. Department of Computer Science and Engineering National Institute of Technology Sikkim Sikkim Ravangla India Department of Computer Application National Institute of Technology Raipur Raipur India Qingdao China Instituto de Telecomunicações Covilhã Portugal

In recent era, the unmanned aerial vehicles (UAVs) commonly known as drones has emerged as a one of the most significant and promising tools which has demonstrated its wide range of implementations variating from commercial domain to the field of defense due to its distinct capabilities such as inspection, surveillance, precision and so forth. Internet connected drones provides a propitious trend that boosts the flying safety, and service qualities of the UAVs where numerous low-altitude drones winged in different flying regions for executing a precise task such as gathering the real-time information from the unuttered environment to be interpretated by users. Nevertheless, the open-access insecure communications in hostile environment, the issues like safety and confidentiality threats, various security concerns such as the leakage of flying courses, identities, position, and gathered data by the drones are upstretched. To address these security concerns, access control mechanism provides a potential service in terms of authentication and key agreement for securing the communication between the individual drones within their respective flying regions. This article introduces a robust, efficient, lightweight, and privacy preserving ECC integrated access control approach by employing digital certificate with considering the high dynamicity and mobility of the drones. The designing of this proposed approach, that is, CLAASC-IoD is influenced by the aim of inter-drone and drone-to-ground station communication in the IoD paradigm. The detailed analysis of security using probabilistic random oracle model as well as simulation using well-accepted security verification tool AVISPA and comparative performance evaluation supports the claim of robustness, effectiveness, and proficiency. © 2023 John Wiley & Sons Ltd.

关键词： Internet of things

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：