检索结果-内蒙古大学图书馆

4th International Conference on Multimedia Processing, Communication and Information technology, MPCIT 2024

作者： Ubale, Ganesh Kashyap, Aryan Kalal, Sahil Kamble, Omkar Pande, Soham Mhamane, Saniya Vishwakarma Institute of Technology Department of Artificial Intelligence and Data Science Pune India

ISBN: (纸本)9798350375466

This study evaluates the performance of pre-trained convolutional neural networks (CNNs) for multi-class skin disease detection. We investigate the efficacy of four models: Xception, InceptionResNetV2, ResNet50, and VGG16. The models are trained and tested on two publicly available datasets: DermNet (6 diseases) and a skin disease classification dataset. We compare the performance of each model based on metrics like accuracy, precision, recall, and F1-score. This research aims to identify the most suitable pre-trained CNN architecture for accurate skin disease classification using readily available datasets. The findings can contribute to the development of automated diagnostic tools to assist dermatologists in early and precise skin disease detection. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Retinal Disease Classification: Binary and Multiclass Perspectives 4

Retinal Disease Classification: Binary and Multiclass Perspe...

引用

4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025

作者： Lanjewar, Rutuja Nitin Pacharaney, Utkarsha Faculty of Engineering and Technology Department of Artificial Intelligence and Data Science Maharashtra Wardha India

ISBN: (纸本)9798331523923

Retinal diseases, among which diabetic retinopathy, age-related macular degeneration, and glaucoma account for the greatest number of blindness cases worldwide, affect more than 338 million people. It is essential that early detection leads to timely intervention to preserve vision, but it is limited due to the ever-increasing demand for ophthalmic care and a shortage of specialists. Machine learning-driven systems, especially Convolutional Neural Networks (CNNs), have shown great promise in overcoming these challenges, achieving high accuracy in medical imaging tasks. This paper explores a unified framework for both multiclass and binary classification of retinal diseases, addressing issues like class imbalance and inter-class similarity while leveraging CNNs for robust feature extraction. Techniques like Grad-CAM also increase model interpretability, allowing clinicians to understand and trust AI-driven diagnostic processes. © 2025 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Decentralized Application using Blockchain 5

Decentralized Application using Blockchain

引用

5th International Conference on Image Processing and Capsule Networks, ICIPCN 2024

作者： Kayade, Parth Pardeshi, Aniket Patil, Suyog Raut, Prashant Shetkar, Pranav Barhate, Madhuri Vishwakarma Institute of Technology Department of Artificial Intelligence and Data Science Pune India

ISBN: (纸本)9798350367171

This research work introduces a novel social media platform built on blockchain technology, aiming to address contemporary challenges in privacy, security, and content integrity. Leveraging decentralized ledger technology, the platform ensures transparent and secure data storage, granting users unprecedented control over their personal information. Through cryptographic protocols, identity protection is prioritized, reducing vulnerability to data breaches. The proposed system employs smart contracts for content attribution, guaranteeing immutable and authentic user-generated content. Token-based incentives incentivize community participation and amplify content creators' contributions. By emphasizing decentralization and transparency, the developed platform intends to revolutionize the social media landscape, fostering a more resilient, userdriven digital environment. The fusion of blockchain technology and social media heralds a new era, offering a resilient, transparent, and user-centric digital ecosystem for the future. © 2024 IEEE.

关键词： Dapp NodeJS ReactJS Smart Contract Solidity Web3.0

来源：评论

学校读者我要写书评

暂无评论

Recent metaheuristic algorithms for medical object localization using MSER detector in computer-aided diagnosis system

引用

Multimedia Tools and Applications 2025年第15期84卷 14433-14487页

作者： Ait Mehdi, Mohamed Belattar, Khadidja Souami, Feryel Faculty of Computer Science Data Science and Artificial Intelligence Department Laboratory for Research in Artificial Intelligence University of Science and Technology Houari Boumediene BP 32 El-Alia Algiers Bab-Ezzouar16111 Algeria Department of Computer Science University of Algiers Benyoucef Benkhedda Algiers16000 Algeria

Object localization is a critical task in image analysis, often facilitated by artificial intelligence techniques. While the Maximally Stable Extremal Regions (MSER) detection algorithm is a popular choice for local detection, its exhaustive approaches can be algorithmically complex and prone to suboptimal results with improper parameter selection. Various metaheuristic algorithms have been proposed for medical object localization to address this. In this context, four contributions are presented. Firstly, recent metaheuristics, including the Slime Mould Algorithm (SMA), Marine Predators Algorithm (MPA), Heap-based Optimizer (HBO), and Gradient-based Optimizer (GBO), are adapted to tackle the MSER localization problem. These algorithms are rigorously evaluated across diverse medical image datasets using multiple metrics, with their performance statistically validated through the Friedman mean rank test. The second contribution introduces a novel objective function to improve the localization process. The third contribution involves a comparative analysis of the recent algorithms against seven standard metaheuristics specifically designed for MSER localization. Lastly, we present an improved computer-aided diagnosis system that integrates an SVM-based model with Local Binary Patterns (LBP) descriptors extracted from each MSER alongside DenseNet-121 features. Experimental results demonstrate that the HBO and SMA optimizers outperform conventional algorithms, showing a higher diversity, improved exploitation and exploration capabilities, a well-balanced exploration-exploitation trade-off, elevated fitness values, faster convergence speeds, and enhanced computational efficiency. Additionally, we integrate these findings into an improved computer-aided diagnosis system, yielding classification results surpassing state-of-the-art models. These findings highlight the potential advantages of this approach, especially in object detection applications. © The Author(s), under

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Analysis of GAN for Melanoma Skin CancerClassification with Dermatologist Recommendation

Analysis of GAN for Melanoma Skin CancerClassification with ...

引用

2024 International Conference on Recent Advances in Electrical, Electronics, Ubiquitous Communication, and Computational intelligence, RAEEUCCI 2024

作者： Saranya, N. Jowin Alfred, C. Rr, Rishikesh Gilbert Idayan, I. KPR Institute of Engineering and Technology Department of Artificial Intelligence and Data Science India

ISBN: (数字)9798350354539

ISBN: (纸本)9798350354539

Melanoma is a type of skin cancer that poses a significant health threat globally, necessitating effective early detection methods. Melanoma arises from the pigment-producing cells(melanocytes) and is known for its higher metastatic potential. Designing a reliable Computer-Aided Diagnosis (CAD) system for accurate melanoma detection is challenging due to the appearance of the nevus. This study investigates the application of a GAN for classification of melanoma. The GAN architecture is employed to learn and generate synthetic melanoma images, enhancing the diversity and richness of the training dataset. We mainly focus on evaluating performance of GAN-enhanced model through comprehensive metrics like specificity, accuracy, the receiver operating characteristic curve (ROC) and area under the curve (AUC). This study sheds light on utility of synthetic data in clinical image analysis and provides insights into the strengths and limitations of utilizing GANs for melanoma skin cancer classification. We also added a dermatologist recommendation feature to our model. It uses computer programs to understand what youneed and then matches you with the best doctor available nearby. This system helps you get the care you need faster and easier, making sure you see a doctor who's a good fit for you and your health needs. This system matches patients with suitable doctors. This matching involves considering factors like doctor specialty and location. © 2024 IEEE.

关键词： Dermatology

来源：评论

学校读者我要写书评

暂无评论

Deep Learning for Automated Image Captioning: A CNN and Transformer Model Analysis 3

Deep Learning for Automated Image Captioning: A CNN and Tran...

引用

3rd International Conference on Intelligent data Communication Technologies and Internet of Things, IDCIoT 2025

作者： Saarika, Kathula Varsha, Vasaraju Sri Harsha, P. Sheikh, Anjum Nabi Dept. Artificial Intelligence and Data Science Chaitanya Bharathi Institute of Technology India

ISBN: (纸本)9798331527549

Automatic image captioning is a multidisciplinary effort that combines machine learning, natural language processing, and computer vision in creating meaningful and context-rich captions for images. Within a short time, this research problem has garnered a lot of attention and new techniques are continuously being proposed. However, there are many dismaying factors as well. For instance, the current models still have a long way to go in attaining human accuracy. This paper seeks to conduct a systematic review of the recent trends and developments in the field of image captioning techniques and models with a focus on 2023 and 2024 by systematically analyzing the 13 publications located in Springer and IEEEXplore libraries. According to our research, Convolutional Neural Networks allow the identification of image content at an early stage, and Recurrent Neural Networks or Long Short-Term Memory networks usually appear at the text generation stage. The most common image datasets are MS COCO and Flickr8k. Almost all studies utilized the BLEU scores (1, 2, 3, and 4) mechanism. It is alleviated that CNN with LSTM models tends to perform better when compared to RNN-modeled networks. Attention mechanisms and Encoder-Decoder architecture are aggressively becoming the two major solutions to strengthening image captioning. This review will help and provide direction to those who are interested in practicing research that aims to promote the techniques and processes of automatic image captioning. © 2025 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Trajectory Tracking Control for Intelligent Wheelchairs Based on Super-Twisting Sliding Mode Observer 5

Trajectory Tracking Control for Intelligent Wheelchairs Base...

引用

5th International Seminar on artificial intelligence, Networking and Information technology, AINIT 2024

作者： Dong, Meiqi Peng, Haojie Zhang, Yan Wu, Jianfei Wang, Zhixiang Yang, Yixin Hebei University of Technology College of Artificial Intelligence and Data Science Tianjin China

ISBN: (纸本)9798350385557

As the rapidly rising number of elders and disabled people in China, intelligent wheelchairs become important to enhance the mobility for them. Many advanced technologies are integrated in intelligent wheelchairs, of which the stable and accurate motion control is the most important due to all the upper functions are depended on this. To solve the problems that the dynamic model of the intelligent wheelchair is difficult to establish accurately, and various disturbances in the movement process cause yaw to the high-precision trajectory tracking, a dual-loop trajectory tracking control method is investigated. Firstly, a virtual velocity controller is developed based on the trajectory tracking error. Then, a super-twisting sliding mode observer (STSMO) is constructed using the dynamic model to estimate the disturbance of the system, and the estimation error converges to zero in finite time. Based on this, a fast integral terminal sliding mode controller (FITSMC) is designed to accurately track the virtual velocity in finite time. Finally, the stability of the dual-loop system is guaranteed by the Lyapunov stability theory. Simulation results demonstrate that the recommended trajectory tracking control method has a faster response speed and stronger robustness. © 2024 IEEE.

关键词： Dynamic models

来源：评论

学校读者我要写书评

暂无评论

Enhancing Summarization of Legal Text Documents using Pre-trained Models 5

Enhancing Summarization of Legal Text Documents using Pre-tr...

引用

5th International Conference on Emerging Systems and Intelligent Computing, ESIC 2025

作者： Kasar, Ashish Matade, Suyash Rasal, Durvankur Shinde, Swapnil Vishwakarma Institute of Information Technology Department of Artificial Intelligence and Data Science Pune India

ISBN: (纸本)9798331522100

A very relevant challenge of summarizing legal text documents often exceeds the length, therefore being complex, was addressed using pre-trained models such as BART and PEGASUS. In the paper, the issue is discussed-the challenge that these pre-trained models face while being effective in general text summarization tasks, which becomes difficult in legal contexts due to specialized vocabulary and structure of legal documents. The authors fine-tune these models using the BillSum dataset-provided US Congressional bills as a proxy for court documents. Evaluation on ROUGE metrics reveals that BART outperforms PEGASUS in summation of content structures of legal texts to the greatest extent intended. The results suggest that the pre-trained models significantly improve the process of summarizing legal documents, both for professionals in law and general readers who would otherwise have to wade through an extensive amount of legal information. © 2025 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Real - Time Speech Decoding through Advanced lip Reading using Deep Learning Model 2

Real - Time Speech Decoding through Advanced lip Reading usi...

引用

2nd IEEE International Conference on Emerging Research in Computational science, ICERCS 2024

作者： Tarun, M. Bairavel, S. Sudha, D. John, Jobin Vignesh, G. KCG College of Technology Department of Artificial Intelligence And Data Science Chennai India

ISBN: (数字)9798331534967

ISBN: (纸本)9798331534950

This research aims to develop a real-time speech decoding system by advanced lip-reading techniques through a deep learning model. The proposed solution integrates cutting-edge advancements in deep learning to make improvement in the efficiency of speech recognition systems. By employing this model will be trained on large datasets of audio-visual speech samples to learn analyzing the patterns of lip movements corresponding to speech phonetics. The system's real-time capability will be achieved through optimized model architectures and efficient computational strategies. The implementation will be validated using diverse datasets and evaluated against standard benchmarks, demonstrating its potential for applications in noisy environments, accessibility for the hearing-impaired, and improved speech recognition in various real-world scenarios. This research contributes to the evolving field of multimodal deep learning, offering a promising avenue for real-time speech decoding through advanced lip reading. © 2024 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

MediScan : A Robust Xception-Based Skin Disease Classification System with Blockchain Integration 3

MediScan : A Robust Xception-Based Skin Disease Classificati...

引用

3rd IEEE International Conference on Blockchain and Distributed Systems Security, ICBDS 2024

作者： Ubale, Ganesh Kashyap, Aryan Kalal, Sahil Kamble, Omkar Pande, Soham Mhamane, Saniya Vishwakarma Institute of Technology Department of Artificial Intelligence and Data Science Pune India

ISBN: (纸本)9798350354348

This research aims to revolutionize the prediction of skin diseases in patients by employing advanced deep learning methodologies. By integrating transfer learning with Xception and Convolutional Neural Network (CNN) models, we develop a robust predictive framework. Transfer learning is leveraged to accelerate training and improve accuracy by using pre-trained models on patient-specific data. To address data security concerns, we incorporate blockchain technology to securely store patient data, ensuring transparency and maintaining integrity. The model's efficacy is demonstrated on a comprehensive skin disease image dataset comprising over 3,396 images across 6 different skin conditions. Results show an improvement in diagnostic accuracy to 85%, a significant enhancement over traditional methods. This integration of transfer learning, advanced CNN models, and blockchain establishes a framework that enhances dermatological diagnostics and ensures a secure, transparent ecosystem for managing patient health data, setting new standards in predictive analytics. © 2024 IEEE.

关键词： Convolutional neural networks

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：