检索结果-内蒙古大学图书馆

2023 International Conference on Evolutionary Algorithms and Soft Computing Techniques, EASCT 2023

作者： Pasha, Syed Nawaz Anitha, D. Anbunathan, R. Gobinath, C. Madhura, G.K. Sr University School of Computer Science & Artificial Intelligence Warangal India School of Computing Srm Institute of Science and Technology Department of Computational Intelligence Kattankulathur India Sa Engineering College Department of Mathematics Chennai India Reva University School of Computing and Information Technology Bangalore India Nitte Meenakshi Institute of Technology Department of Ai and Ml Bengaluru India

ISBN: (纸本)9798350313413

Iris recognition stands out as one of the most stable and accurate biometric identification systems such as facial, voice, fingerprint, handwriting, and iris recognition itself. Deep Learning (DL) plays a pivotal role in enhancing security by leveraging Convolutional Neural Networks (CNNs) for iris recognition. The acquisition of iris images and real-time processing have become focal points demanding heightened attention in the realm of DL. Iris recognition is prized for its high level of protection, making it challenging to reconstruct iris data easily from the database. The CASIA-Iris Dataset serves as a benchmark to verify the uniqueness of iris patterns and evaluate the performance of iris recognition. However, due to limitations in iris image access, the application of CNNs in the segmentation approach remains critical for iris localization and pupil boundary identification, ensuring both efficiency and specificity in iris images. The results obtained from this study reveal that the proposed CNN, utilizing DL models, achieves an impressive accuracy of 99.89% on the CASIA-Iris dataset, better performance of other existing models, such as U-Net and VGG-Net, and ensuring accurate classification. © 2023 IEEE.

关键词： Contrast-Limited Adaptive Histogram Equalization Convolution Neural Network Deep Learning Iris Recognition and Support Vector Machine

来源：评论

学校读者我要写书评

暂无评论

Colorization of Grayscale Images Using Neural Network Without User Intervention

Colorization of Grayscale Images Using Neural Network Withou...

引用

Advancement in Electronics & Communication engineering (AECE), International Conference on

作者： Vinish Kumar Santosh Verma Gaurav Raj Department of Computer Science and Engineering (AI&ML) RKGIT Ghaziabad UP India Course Mentor-Geeksforgeeks Noida UP India

The task of colorizing grayscale images has been performed manually for a long time. Recently, neural networks have found their use in this domain. In this project we propose a system that takes a grayscale image as an input and produces a colored image of the same. The technique implemented in this project does not require complicated human inputs or image processing techniques, but rather training of the neural network much like a human mind. Where manual colorization methods can take months to color a single image, neural networks, once trained, are much more efficient in terms of time consumption and also comparable levels of accuracy can be achieved which depends upon training of the neural network. The Neural network works on hit and trial method; at first it makes a random prediction for a pixel, calculates the error and then tries to reduce it as much as possible by the method of backpropagation. The classifier, trained using backpropagation of information, predicts the color of a pixel based on the grayscale value of its own as well as the surrounding pixels in the grayscale image.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Glioma Detection using Improved Artificial Neural Network in MRI Images 2

Glioma Detection using Improved Artificial Neural Network in...

引用

2nd IEEE Mysore Sub Section International Conference, MysuruCon 2022

作者： Nandihal, Priya Shetty S, Vijaya Guha, Tapas Pareek, Piyush Kumar Global Academy of Technology Department of AI/ML Bengaluru India Nitte Meenakshi Institute of Technology Department of Computer Science Engineering Bangalore India Nitte Meenakshi Institute of Technology Department of Artificial Intelligence & Machine Learning Bengaluru India

ISBN: (数字)9781665497909

ISBN: (纸本)9781665497909

The physical identification of tumors may be a laborious and time-consuming process for medical professionals because of the complex nature of the tumor and the noise involution that can occur in magnetic resonance (MR) imaging information. Therefore, determining the location of the tumor at an earlier stage is quite important. The medical scan may track and prognosticate the uncontrolled proliferation of cancer pretentious regions at different levels in order to deliver a felicitous diagnosis at an early time. This is accomplished via the utilisation of segmentation in conjunction with relegation procedures. In order to recognize the tissues of a brain tumor, segmentation of the picture obtained from the MRI is a crucial and challenging step. So, the proposed work includes the tumor segmentation process using Region Growth Algorithm with Gray-Level-Run-Length-Matrix and Centre-Symmetric-Local-Binary-Patterns texture feature extraction process. The segmented images undergo feature extraction process with higher level of accuracy. The performance metrics are measured using accuracy, sensitivity and specificity. The proposed work has 0.97% sensitivity, 0.85% specificity and 99.80% accuracy. © 2022 IEEE.

关键词： Tumors

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Based Smart Contracts for Supply Chain Management: A Strategic Analysis 3

Blockchain-Based Smart Contracts for Supply Chain Management...

引用

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

作者： Santosh, Kathari Dash, Chinmaya Aarif, Mohd Kumar, Yadavalli Suresh Verma, Mrinal Kiranbala, B. Cmr Institute of Technology Department of Mba Bengaluru India St. Claret College Department of Computer Science Bangalore India Global Research Network Noida India Aditya College of Engineering Department of Computer Science & Engineering Andhra Pradesh India Soil School of Business Design Manesar Gurgaon India K.Ramakrishnan College of Engineering Department of Ai and Ds Trichy India

ISBN: (纸本)9798350361186

The supply chain management sector strategic examination of the blockchain relies smart contract technology employs a comprehensive approach, drawing insights from reputable secondary sources to explore performance metrics, transactional flow optimization, consensus mechanisms, and security considerations. The study reveals nuanced performance variations in average throughput and transaction duration, emphasizing the scalability and stability of the blockchain network. The depicted transactional flow showcases enhanced communication and transparency, aligning with strategic goals for supply chain optimization. The introduction of a Proof of Authority (PoA) consensus algorithm underscores the importance of efficient consensus mechanisms for resilience against attacks. By integrating PoA, the proposed blockchain-based smart contracts contribute to a strategic and reliable supply chain management system, enhancing transparency and mitigating potential risks. A robust security analysis demonstrates moderate fault tolerance, high defense mechanisms, and effective security recovery capabilities, aligning with the overarching goals of confidentiality, integrity, and availability of critical supply chain data. Ethereum's Proof of Stake (PoS) mechanism is referenced as an exemplar for improving scalability and security. The comparative analysis highlights the proposed CMBC (Consensus Mechanism with Blockchain) as a standout solution, offering middle-ground high attack diversity, fault tolerance, recovery, and a high attack cost, robust security. CMBC outperforms alternatives, demonstrating superiority in enhancing security and resilience against potential adversarial threats. The proposed CMBC demonstrates notable strengths across key performance indices when compared to Fortified-Chain and SaYoPillow. The proposed method is implemented using python software. The effective standard deviation values, ranging from 9.58 to 25.13, further underscore the system's stability and re

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Face Detection in Video Live Stream Using Deep Learning 2

Face Detection in Video Live Stream Using Deep Learning

引用

2nd International Conference on Disruptive Technologies, ICDT 2024

作者： Amrita Kumar, Rajiv Joshi, Shivani Raj, Daksh Centre for Cyber Security and Cryptology Computer Science & Engineering Sharda School of Engineering & Technology Sharda University U.P. Greater Noida India G L Bajaj Institute of Technology and Management Department of Applied Computational Science and Engineering Nvidia Ai Learning & Research Center U.P. Greater Noida India G L Bajaj Institute of Technology and Management Department of Computer Science and Engineering U.P. Greater Noida India Galgotias College of Engineering and Technology Department of Computer Science and Engineering U.P. Greater Noida India

ISBN: (纸本)9798350371055

With the increasing popularity of live streaming platforms, there is an imperative need for robust methods for identifying individuals in real-time. The proposed system utilizes Convolutional Neural Networks (CNN) computer vision algorithms to identify faces in live streams and match them against a pre-existing database of known individuals. Additionally, the system employs deep learning and supervised learning techniques along with various types of CNN to achieve robust facial recognition in diverse lighting and environmental conditions. The system's accuracy is evaluated using various performance metrics, including precision, recall, and F1-score. The results demonstrate that the proposed system achieves high accuracy in person identification during real-time live streams. This system holds potential applications in a variety of domains, including security, entertainment, and social media. © 2024 IEEE.

关键词： computer vision Convolutional neural networks (CNN) deep learning F1-score Live streaming precision recall supervised learning

来源：评论

学校读者我要写书评

暂无评论

Exploratory Diffusion Policy for Unsupervised Reinforcement Learning

arXiv

引用

arXiv 2025年

作者： Ying, Chengyang Chen, Huayu Zhou, Xinning Hao, Zhongkai Su, Hang Zhu, Jun Department of Computer Science & Technology Institute for AI BNRist Center Tsinghua-Bosch Joint ML Center THBI Lab Tsinghua University China

Unsupervised reinforcement learning (RL) aims to pre-train agents by exploring states or skills in reward-free environments, facilitating the adaptation to downstream tasks. However, existing methods often overlook the fitting ability of pre-trained policies and struggle to handle the heterogeneous pre-training data, which are crucial for achieving efficient exploration and fast fine-tuning. To address this gap, we propose Exploratory Diffusion Policy (EDP), which leverages the strong expressive ability of diffusion models to fit the explored data, both boosting exploration and obtaining an efficient initialization for downstream tasks. Specifically, we estimate the distribution of collected data in the replay buffer with the diffusion policy and propose a score intrinsic reward, encouraging the agent to explore unseen states. For fine-tuning the pre-trained diffusion policy on downstream tasks, we provide both theoretical analyses and practical algorithms, including an alternating method of Q function optimization and diffusion policy distillation. Extensive experiments demonstrate the effectiveness of EDP in efficient exploration during pre-training and fast adaptation during fine-tuning. Copyright © 2025, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Domain Aware Multi-task Pretraining of 3D Swin Transformer for T1-Weighted Brain MRI 17th

Domain Aware Multi-task Pretraining of 3D Swin Transformer ...

引用

17th Asian Conference on computer Vision, ACCV 2024

作者： Kim, Jonghun Kim, Mansu Park, Hyunjin Department of Electrical and Computer Engineering Sungkyunkwan University Suwon Korea Republic of Center for Neuroscience Imaging Research Institute for Basic Science Suwon Korea Republic of AI Graduate School Gwangju Institute of Science and Technology Gwangju Korea Republic of

ISBN: (纸本)9789819609000

The scarcity of annotated medical images is a major bottleneck in developing learning models for medical image analysis. Hence, recent studies have focused on pretrained models with fewer annotation requirements that can be fine-tuned for various downstream tasks. However, existing approaches are mainly 3D adaptions of 2D approaches ill-suited for 3D medical imaging data. Motivated by this gap, we propose novel domain-aware multi-task learning tasks to pretrain a 3D Swin Transformer for brain magnetic resonance imaging (MRI). Our method considers the domain knowledge in brain MRI by incorporating brain anatomy and morphology as well as standard pretext tasks adapted for 3D imaging in a contrastive learning setting. We pretrain our model using large-scale brain MRI data of 13,687 samples spanning several large-scale databases. Our method outperforms existing supervised and self-supervised methods in three downstream tasks of Alzheimer’s disease classification, Parkinson’s disease classification, and age prediction tasks. The ablation study of the proposed pretext tasks shows the effectiveness of our pretext tasks. Our code is available at ***/jongdory/DAMT. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Neurodegenerative diseases

来源：评论

学校读者我要写书评

暂无评论

Comment Sentiment Analysis Using Bidirectional Encoder Representations from Transformers

SSRN

引用

SSRN 2024年

作者： Mittal, Harshit Garg, Neeraj Computer Science Department Maharaja Agrasen Institute of Technology Delhi India AI/ML Department Maharaja Agrasen Institute of Technology Delhi India

This research paper unravels the power of Bidirectional Encoder Representations from Transformers (BERT), a cutting-edge language representation model, in the realm of comment sentiment analysis. By focusing on two main aspects - the working principles of BERT and the methodology of comment sentiment analysis - the paper aims to captivate readers with an engaging exploration of this exciting field. In this paper, we delve into the remarkable workings of BERT, showcasing its exceptional ability to grasp context, semantics, and sentiment in natural language text. Through a comprehensive examination of its underlying architecture, intricate pre-training process, and ingenious fine-tuning techniques, readers will gain a profound understanding of how BERT transforms raw language into rich, contextualized representations. Building upon this foundation, the paper unveils the methodology employed in our comment sentiment analysis project. Here, we take readers on an enthralling journey through the data collection and pre-processing phases, carefully selecting appropriate evaluation metrics, and ingeniously designing a sentiment classification framework. Along the way, we address the challenges of word sense disambiguation and tackle the nuances of neutral sentiments. By venturing into the uncharted territory of comment sentiment analysis using BERT, this research paper becomes more than just an invaluable resource for practitioners and researchers. It serves as a seductive invitation to unlock the hidden treasures within user opinions and sentiments, enabling us to leverage this invaluable wealth of information across various domains. With insightful findings and actionable recommendations, this paper fuels the pursuit of advanced sentiment analysis techniques, propelling the field towards enhanced comprehension and optimal utilization of user sentiments in the ever-evolving digital landscape. © 2024, The Authors. All rights reserved.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Exploring Diabetic Retinopathy Diagnosis: Insights from a Comprehensive Literature Review and Initial Fundus Image Analysis

Exploring Diabetic Retinopathy Diagnosis: Insights from a Co...

引用

2024 International Conference on Electrical, Electronics and Computing Technologies, ICEECT 2024

作者： Kumar Reddy, K Uday Malleswari, M. Kumar Reddy, P Phanindra Karimullah, Shaik Shaik, Fahimuddin Hariobulesu, Peddapullaiahgari Annamacharya University Artificial Intelligence and Data Science Department Rajampet India Sri Annamacharya Institute of Technology and Sciences Computer Science and Engineering Department India Annamacharya University AI&DS Department Rajampet India Annamacharya University ECE Department Rajampet India

ISBN: (纸本)9798350378092

This article looks at recent advances in diabetic retinopathy (DR) detection with retinal fundus and optical coherence tomography angiography (OCTA) pictures. Key studies have investigated a variety of methods, including combining Multiscale Retinex with Chromaticity Preservation (MSRCP) and Bi-LSTM networks for high accuracy, integrating the Internet of Things (IoT) with deep learning for increased sensitivity, and using Multi-Stream Deep Neural Networks to classify DR severity. Innovations like the PMNPDR system for microaneurysm detection and deep learning ensembles for OCT images have also performed well. Despite these gains, issues like as data transfer latency, model complexity, and reliance on tiny datasets remain. Overall, while current models show great potential, they require further modification and validation to assure their efficacy and applicability in clinical situations. © 2024 IEEE.

关键词： Optical coherence tomography

来源：评论

学校读者我要写书评

暂无评论

Converging to a Lingua Franca: Evolution of Linguistic Regions and Semantics Alignment in Multilingual Large Language Models 31

Converging to a Lingua Franca: Evolution of Linguistic Regio...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Zeng, Hongchuan Han, Senyu Chen, Lu Yu, Kai X-LANCE Lab Department of Computer Science and Engineering MoE Key Lab of Artificial Intelligence SJTU AI Institute Shanghai Jiao Tong University Shanghai China Suzhou Laboratory Suzhou China

ISBN: (纸本)9798891761964

Large language models (LLMs) have demonstrated remarkable performance, particularly in multilingual contexts. While recent studies suggest that LLMs can transfer skills learned in one language to others, the internal mechanisms behind this ability remain unclear. We observed that the neuron activation patterns of LLMs exhibit similarities when processing the same language, revealing the existence and location of key linguistic regions. Additionally, we found that neuron activation patterns are similar when processing sentences with the same semantic meaning in different languages. This indicates that LLMs map semantically identical inputs from different languages into a "Lingua Franca", a common semantic latent space that allows for consistent processing across languages. This semantic alignment becomes more pronounced with training and increased model size, resulting in a more language-agnostic activation pattern. Moreover, we found that key linguistic neurons are concentrated in the first and last layers of LLMs, becoming denser in the first layers as training progresses. Experiments on BLOOM and LLaMA2 support these findings, highlighting the structural evolution of multilingual LLMs during training and scaling up. This paper provides insights into the internal workings of LLMs, offering a foundation for future improvements in their cross-lingual capabilities. The codes are available at: https://***/X-LANCE/LinguaFranca. © 2025 Association for Computational Linguistics.

关键词： Neurons

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：