检索结果-内蒙古大学图书馆

3rd IEEE Asian Conference on Innovation in technology, ASIANCON 2023

作者： Gautam, Vinay Tiwari, Raj Gaurang Trivedi, Naresh Kumar Jain, Anuj Kumar Sharma, Vikrant Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Graphic Era Hill University Deparment of Computer Science and Engineering Dehradun India

ISBN: (纸本)9798350302288

Several classification problems have been resolved with computer vision and deep learning technology. Despite this, no dedicated work has been carried out to classify plastic bottles from non-bottle objects. Therefore, this paper is concerned to resolve issue to classify waste plastic bottle from non-bottle objects. Here a dataset of 52000 images of bottle and non-bottle has been used to classify different objects based on their size. Each image contains a single bottle and non-bottle object. In this paper, various pre-trained models have been employed to classify these objects based on their size. The pretrained models have been trained with bottle dataset and attained 97%, 92% and 91 % accuracy. The comparison shows that the ResNet-18 outperforms. © 2023 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Approach for Marine Animal Classification: Enhancing Taxonomic Identification and Conservation Efforts 3

A Deep Learning Approach for Marine Animal Classification: E...

引用

3rd IEEE Asian Conference on Innovation in technology, ASIANCON 2023

作者： Tiwari, Raj Gaurang Gautam, Vinay Trivedi, Naresh Kumar Jain, Anuj Kumar Sharma, Vikrant Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Graphic Era Hill University Deparment of Computer Science and Engineering Dehradun India

ISBN: (纸本)9798350302288

Classification of marine animals is essential for studying marine ecosystems, keeping track of biodiversity, and developing efficient conservation plans. The success of deep learning methods in picture classification tasks has increased dramatically in recent years. In this research, we propose a systematic investigation of the use of several deep learning models for the categorization of marine organisms;these models include Inception, ResNetV, ResNet50, VGG16, InceptionV3, Xception, DenseNet, MobileNet, and NASNet, among others. Models originally built on ImageNet were then fine-tuned using transfer learning before being applied to our marine animal dataset for training. Training and validation accuracy were used to rank the models, and the models' overall performance was compared to find the best one. Inception outperformed the other models in tests by a wide margin, with a training accuracy of 95.85% and a validation accuracy of 96.75. Inception's greater performance in marine animal categorization may be attributed to its deeper design and its ability to grasp complex spatial connections. When deep learning models are used to correctly categorize marine species, it may greatly improve taxonomic identification, aid conservation efforts, and make ecological study much easier. Additionally, the results highlight the significance of constructing domain-specific deep learning models for fields like marine biology, where distinct obstacles exist. © 2023 IEEE.

关键词： Image recognition

来源：评论

学校读者我要写书评

暂无评论

Preserving India's Rich Dance Heritage: A Classification of Indian Dance Forms and Innovative Digital Management Solutions for Cultural Heritage Conservation

Preserving India's Rich Dance Heritage: A Classification of ...

引用

2023 International Research Conference on Smart Computing and Systems Engineering, SCSE 2023

作者： Tiwari, Raj Gaurang Gautam, Vinay Sharma, Vikrant Jain, Anuj Kumar Trivedi, Naresh Kumar Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Graphic Era Hill University Deparment of Computer Science and Engineering Dehradun India

ISBN: (纸本)9798350341454

Deep connections exist between dance and cultural heritage. Dance is frequently passed down through generations as an essential component of a culture's identity and as a means of maintaining and honoring that culture's distinctive traditions and customs. A culture's history, beliefs, and values can be powerfully expressed via dance, which can also be used for communication and storytelling. For the purpose of protecting and promoting India's cultural legacy, it is crucial to comprehend how Indian dance styles are categorized. This paper proposes a hybrid Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) deep learning approach for the accurate classification of Indian dance style categories. The proposed model combines the strengths of both CNN and RNN to leverage spatial and temporal information, respectively, resulting in enhanced performance and improved accuracy. Extensive experiments were conducted to evaluate the performance of the proposed approach. The results demonstrate that the hybrid CNN-RNN model achieved an impressive accuracy of 97.74%, outperforming traditional methods and single-model architectures. © 2023 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Does Image Anonymization Impact computer Vision Training?

arXiv

引用

arXiv 2023年

作者： Hukkelås, Håkon Lindseth, Frank Deparment of Computer Science Norwegian University of Science and Technology Norway

Image anonymization is widely adapted in practice to comply with privacy regulations in many regions. However, anonymization often degrades the quality of the data, reducing its utility for computer vision development. In this paper, we investigate the impact of image anonymization for training computer vision models on key computer vision tasks (detection, instance segmentation, and pose estimation). Specifically, we benchmark the recognition drop on common detection datasets, where we evaluate both traditional and realistic anonymization for faces and full bodies. Our comprehensive experiments reflect that traditional image anonymization substantially impacts final model performance, particularly when anonymizing the full body. Furthermore, we find that realistic anonymization can mitigate this decrease in performance, where our experiments reflect a minimal performance drop for face anonymization. Our study demonstrates that realistic anonymization can enable privacy-preserving computer vision development with minimal performance degradation across a range of important computer vision benchmarks. © 2023, CC BY-SA.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

Ethiopic Cryptography: An Efficient Hybrid Encryption Algorithm for Reducing Cybersecurity Attack 11th

Ethiopic Cryptography: An Efficient Hybrid Encryption Algori...

引用

11th EAI International Conference on Advancement of science and technology, ICAST 2023

作者： Akele, Tadesse Birara Bahiru, Tadesse Kebede Molla, Tibebu Legesse Fentie, Shambel Dessale Belay, Tadesse Destaw Marye, Almaz Yemataw Getaneh, Mersha Derese Department of Information System Kombolcha Institute of Technology Wollo University Kombolcha Ethiopia Department of Information Technology Kombolcha Institute of Technology Wollo University Kombolcha Ethiopia Deparment of Software Engineering Kombolcha Institute of Technology Wollo University Kombolcha Ethiopia Department of Computer Science Kombolcha Institute of Technology Wollo University Kombolcha Ethiopia

ISBN: (纸本)9783031641503

Cryptography is an essential tool in ensuring the security of data during network communication. Network and service providers prioritize data security to ensure secure data exchanges. Ethiopic cryptography is a unique encryption method that utilizes the Amharic language, prevalent in Ethiopia. Unlike Latin alphabets, Amharic consists of 287 characters, making decryption challenging. The encryption process involves converting the text to Amharic and then mapping it to a pair of randomly chosen two-bit English alphabets, creating an intermediate cipher. The Ethiopic cryptography algorithm employs a hybrid approach, using the RSA algorithm for secure key exchange and the Advanced Encryption Standard algorithm for text confidentiality. The algorithm comprises three phases—translation, mapping, and encryption—contributing to its effectiveness in securing data over the network. Decryption requires familiarity with the Amharic language and mapping details. The algorithm demonstrates a stronger avalanche effect (97%) in comparison with Blowfish and DES. The algorithm is proposed and evaluated using the design science research methodology, and its performance is validated through simulation using MATLAB and BunnaScript tools, considering factors like the avalanche effect, time complexity, and memory consumption. The evaluation shows that the suggested solution operates faster, requires lesser time for data encryption, and uses less memory than other methods in use. Overall, the proposed Ethiopic cryptography algorithm proves to be an efficient and effective method for securing data during network communication, offering stronger text confidentiality and an improved avalanche effect. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： MATLAB

来源：评论

学校读者我要写书评

暂无评论

The Novel Protocol Uses Low Computational Complexity to Authenticate Edge Devices on Smart Grid Systems 9

The Novel Protocol Uses Low Computational Complexity to Auth...

引用

9th IEEE India International Conference on Power Electronics, IICPE 2023

作者： Kuppuswamy, Prakash Al Khalidi, Sayed Q.Y. Kumar, Purushottam Meeran, Ahmed Ali Shaik Sonia Varshini, Vijaya Computer Science Engineering SRM University Haryana Delhi-NCR India King Khalid University Department of Information Technology Abha Saudi Arabia Jazan University Department of Information Technology Jazan Saudi Arabia Bharathiyar University Computer Science Deparment Tamil Nadu Coimbatore India

ISBN: (纸本)9798350307252

Grid computing has emerged as a powerful paradigm for harnessing distributed computing resources to solve complex problems. In grid computing, various devices across different administrative domains collaborate to provide computational power and storage capacity. Sensor agent devices have become the backbone of modern technological systems such as Internet of Things (IoT), smart homes, industrial automation, and autonomous vehicles. These devices collect and process real-time data from the surrounding environment, providing crucial insights for decision-making, monitoring, and control purposes. However, the performance of these devices depends on space complexity and time complexity. Achieving optimum performance in sensor agent devices requires a delicate balance between space complexity and time complexity. Reducing one complexity may increase the other, and vice versa. Light weight authentication device-based algorithms play a vital role in verifying the identity of devices and ensuring secure communication within the grid and that needs to satisfy complexity. These protocols verify the identity of devices and users, ensuring that only authorized entities can access and interact with the data. In this article, we will explore a new light weight authentication protocol based on a block cipher i.e ELBC (Extended Linear Block cipher algorithm). The new authentication protocols are easy to implement and deploy in any kind of environment. The algorithm divides the authentication message into blocks and encrypts them using matrix multiplication. By generating an invertible key matrix and applying encryption and decryption equations, the proposed algorithm ensures secure communication. © 2023 IEEE.

关键词： Authentication

来源：评论

学校读者我要写书评

暂无评论

Progression of Disease Forecasting in Recurrence Cervical Cancer Using Logistic Regression 7

Progression of Disease Forecasting in Recurrence Cervical Ca...

引用

7th International Conference on Electronics, Communication and Aerospace technology, ICECA 2023

作者： Geeitha, S. Renuka, P. Ramya, S. Sinduja, K. M.Kumarasamy College of Engineering Department of Information Technology Karur India Paavai Engineering College Deparment of Computer Science and Engineering Namakkal India Sri Krishna College of Engineering and Technology Department of Information Technology Coimbatore India

ISBN: (纸本)9798350340600

Cervical cancer is the fourth most prevalent disease in women globally as well as main cause of cancer- related mortality among women. The overall survival rates of cervical carcinoma patients are significantly impacted by the detection of recurrence. The disease is entirely treatable, though, if it is found in its early, precancerous levels. According to current figures, the death percentage a few decades ago was excessive. Women who have previously had treatment with CIN3 are most at riskof developing cervical cancer once more. The dataset includes information on the effects of various treatments on post- recurrence survival. The progression of a disease is detected in this research. Of the 4913 patients who were diagnosed withcervical cancer overall, 260 women with the disease suffered recurrences. The Data has been roughly categorized into 27 characteristics that are important for illness staging and therapy. Thevery important attributes selected are FIGO 2009, chemotherapy,radiotherapy, recurrence confirmed by a pathologist, etc. Five different models are trained to predict the progression. Logistic regression is finally predicted with a high level of accuracy of 96.3%. Two hundred and sixty patients with 27 attributes consistof a number of missing values. Research is implemented through the Orange platform. The missing values are balanced through Average and Most frequent data imputation method. Important attributes are selected through the ReliefF technique, and then disease progression is predicted by 5 different algorithms Logistic regression, Support vector machine, Naive Bayes, Stochastic gradient decent and neural network. Finally logistic regression predict the progression of a disease with highest accuracy of 96.3%. © 2023 IEEE.

关键词： Radiotherapy

来源：评论

学校读者我要写书评

暂无评论

State-of-the-Art in Multi-faceted Feature Matching for E-Commerce: A Comprehensive Analysis

Informatica (Slovenia)

引用

Informatica (Slovenia) 2024年第12期48卷 97-106页

作者： Zareena, Noorbasha Rao, B. Tarakeswara Department of Computer Science and Systems Engineering AUTDR HUB Andhra University India Deparment of Computer Science and Engineering Kallam Haranadhareddy Institute of Technology Andhra Pradesh Guntur India

In this paper we provided an insightful exploration into the critical role of feature matching in enhancing the efficacy of e-commerce recommendation systems. By meticulously analyzing user data and product characteristics, feature matching significantly improves the personalization of product suggestions, thereby augmenting user experience and engagement in online shopping platforms. The paper reviews existing methodologies, including collaborative filtering, content-based filtering, hybrid models, and advanced deep learning approaches, which have collectively advanced the domain of personalized recommendations. Despite these advancements, the paper identifies key limitations such as scalability challenges, the cold start problem, over-specialization versus diversity in recommendations, and concerns over user privacy and data security. Through this comprehensive analysis, the survey makes substantial contributions by delineating the current state of the art, pinpointing critical gaps, and suggesting avenues for future research. The paper particularly emphasizes the need for dynamic adaptation in recommendation systems to keep pace with the fast-evolving e-commerce sector and changing consumer behaviors. Future research directions highlighted include the exploration of unsupervised and reinforcement learning models, the development of cross-domain recommendation systems, and the integration of emerging technologies, all aimed at fostering more nuanced, adaptable, and ethically responsible recommendation mechanisms. © 2024 Slovene Society Informatika. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Fundamental Spatial Parameters of Human Mobility

Fundamental Spatial Parameters of Human Mobility

引用

computer science, computer Engineering, & Applied Computing (CSCE), Congress in

作者： Sasha Knezevic Ivica Kostanic Zaid Matloub Deparment of Computer Science and Engineering Florida Institute of Technology Melbourne FL USA

A set of parameters for characterizing spatial aspects of human mobility are presented. The parameters are extracted from a large set of crowd-sourced cellular data which were collected in a major metropolitan area of United States. Each parameter is analyzed at three different time scales (daily, weekly, and monthly time scale). The results demonstrate that human mobility on a population scale has a high degree of statistical regularity. The mobility may be described to a great extend using a limited number of judiciously selected parameters.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Detection and Classification of Rheumatoid Nodule using Deep Learning Models

Detection and Classification of Rheumatoid Nodule using Deep...

引用

2022 International Conference on Machine Learning and Data Engineering, ICMLDE 2022

作者： Dulhare, Uma N. Mubeen, Azmath Department of Computer Science and Artificial Intelligence Muffakham Jah College of Engineering and Technology Hyderabad500034 India Deparment of Computer Science and Engineering University College of Engineering Osmania University Hyderabad500007 India

In the past few years there have been many cases in which the patient may get some form of nodule on the body and it becomes very difficult to identify the type of nodule and its relative disease. Sometimes the nodule may be a simple skin lesion or it may be cancerous cyst or it may be due to a severe condition of rheumatoid arthritis. This research work concentrates on classifying the nodules from a given data set of images and detecting the Rheumatoid Nodule (RN) and its risks. Many studies have been conducted for detection of different nodules but specific to rheumatoid nodule identification and its risks on other organs are minimum as per my cognizance. In imaging study, Deep Convolutional Neural Networks (DCNN) have shown efficient accomplishment, hence the model proposed uses dual three-dimensional Deep Convolutional Networks design for the detection of RN and an improved Res Net for classification, respectively. Rheumatoid Nodule detection was performed through Pure Deep Convolutional Networks (PDCNN) and then feature selection was done using Particle Swarm Optimization (PSO) and later classification was done using the Residual Network (Res-Net) on proficiently learned features from Particle Swarm Optimization. The Deep Learning (DL) model for Rheumatoid Nodule detection and classification, blended with medical aspects, aids in minimizing the errors and False Positive (FP) rate and help in early detection of rheumatoid nodule. The recommended system was assessed on dataset collected from different orthopaedic hospitals in the city and improved results were obtained in comparison to the prevailing approaches. © 2022 Elsevier B.V.. All rights reserved.

关键词： Classification (of information)

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：