检索结果-内蒙古大学图书馆

Multi-View Test-Time Adaptation for Semantic Segmentation in Clinical Cataract Surgery

IEEE Transactions on Medical Imaging 2025年第5期44卷 2307-2318页

作者： Li, Heng Ou, Mingyang Li, Haojin Qiu, Zhongxi Niu, Ke Fu, Huazhu Liu, Jiang Southern University of Science and Technology Research Institute of Trustworthy Autonomous Systems Shenzhen518055 China Southern University of Science and Technology Department of Computer Science and Engineering Shenzhen518055 China Beijing Information Science and Technology University Computer School Beijing100192 China Institute of High Performance Computing Agency for Science Technology and Research Fusionopolis Way 138632 Singapore Southern University of Science and Technology Research Institute of Trustworthy Autonomous Systems Department of Computer Science and Engineering Shenzhen518055 China

Cataract surgery, a widely performed operation worldwide, is incorporating semantic segmentation to advance computer-assisted intervention. However, the tissue appearance and illumination in cataract surgery often differ among clinical centers, intensifying the issue of domain shifts. While domain adaptation offers remedies to the shifts, the necessity for data centralization raises additional privacy concerns. To overcome these challenges, we propose a Multi-view Test-time Adaptation algorithm (MUTA) to segment cataract surgical scenes, which leverages multi-view learning to enhance model training within the source domain and model adaptation within the target domain. In the training phase, the segmentation model is equipped with multi-view decoders to boost its robustness against variations in cataract surgery. During the inference phase, test-time adaptation is implemented using multi-view knowledge distillation, enabling model updates in clinics without data centralization or privacy concerns. We conducted experiments in a simulated cross-center scenario using several cataract surgery datasets to evaluate the effectiveness of MUTA. Through comparisons and investigations, we have validated that MUTA effectively learns a robust source model and adapts the model to target data during the practical inference phase. © 1982-2012 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Improving graph collaborative filtering with network motifs

引用

Neural computing and Applications 2025年 1-20页

作者： Zhang, Yuqi Yu, Jian Liu, Zhizhong Wang, Guiling Nguyen, Minh Sheng, Quan Z. Wang, Nancy Department of Computer and Information Sciences Auckland University of Technology Auckland Central1010 New Zealand School of Computer and Control Engineering Yantai University Shandong Yantai264005 China School of Information Science and Technology North China University of Technology Beijing100144 China School of Computing Macquarie University SydneyNSW2109 Australia

Deep learning on graphs, specifically graph convolutional networks (GCNs), has exhibited exceptional efficacy in the domain of recommender systems. Most GCNs have a message-passing architecture that enables nodes to aggregate information from neighbours iteratively through multiple layers. This enables GCNs to learn from higher-order information, but the model does not allow for direct captions of the local structural patterns. Our rationale is to investigate the effectiveness of capturing such local patterns for graph-based collaborative filtering to enhance model’s learning ability per layer. This technique combines lower-order and higher-order interactions during layer-wise propagation. In this paper, we propose MotifGCN to aggregate both lower-order and higher-order information in each graph convolution layer. Specifically, we develop dedicated algorithms of generating motif adjacency matrices. The matrices are then used for motif-enhanced neighbourhood aggregation in each layer. As this paper focuses on recommender systems, MotifGCN is built on the basis of bipartite graphs. Our experiments on four real-world datasets show that MotifGCN has a superior performance compared to various state-of-the-art methods. © The Author(s) 2025.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

Deep ensemble model for skin cancer classification with improved feature set

引用

Multimedia Tools and Applications 2025年第10期84卷 7599-7626页

作者： Burada, Sreedhar Manjunathswamy, B.E. Kumar, M. Sunil Department of Computer Science and Engineering Don Bosco Institute of Technology Karnataka Bengaluru560074 India Department of Computer Science and Engineering School of Computing Mohan Babu University Sree Vidyanikethan Engineering College Andhra Pradesh Tirupati517102 India

Early identification of skin cancer is mandatory to minimize the worldwide death rate as this disease is covering more than 30% of mortality rates in young and adults. Researchers are in the move of proposing advanced models for the accurate classification of skin disease. This paper introduces a novel skin cancer detection model under the stages of preprocessing, feature extraction, classification as well as segmentation. The given input is first preprocessed via Gaussian filtering. Subsequently, the Deep learning model termed as ‘U-net’ is used for the segmentation process. Afterward, the extraction of the feature set including features like Improved Center symmetric Local Binary Pattern (ICS-LBP), Local Gradient Pattern (LGP), Median binary pattern (MBP) is performed. Classification takes place using Deep Ensemble Model based on the extracted features, that includes BiGRU, Deep Maxout, and Improved Deep Convolutional Neural Network (CNN), respectively. Also, as the hyperparameter tuning of the classification model enhances the performance, this work introduces a new Hybrid Atom and Arithmetic Operation Algorithm (HAAOA) algorithm for tuning the optimal weights of the Bi-GRU (Bi-directional Gated Recurrent Unit) and Deep Maxout model. At last, the validation of the proposed work is performed through different performance measures by comparing it with different algorithms. Accordingly, accuracy of the adopted method at 90th learning percentage is 93.56%, which is (~) 82.63%, 85.15%, 84.98%, 83.66% and 83.42% better than AOA, ASO, BWO, BMO and DO, respectively. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Role-Specific Reward Design with Large Language Model for StarCraft II

Role-Specific Reward Design with Large Language Model for St...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Li, Sijia Luo, Haonan Zhang, Xu Zeng, Xin Shen, Zhixuan Li, Tianrui School of Computing and Artificial Intelligence Southwest Jiaotong University Chengdu China Nanjing University of Science and Technology School of Computer Science and Engineering China

ISBN: (纸本)9798350368741

Reward acts as a signal to guide the agent's learning process in Reinforcement Learning (RL), evaluating and assigning rewards to the agent's actions based on their alignment with goals. Designing reward is challenging in multi-agent environment such as StarCraft II benchmark since agents face credit allocation and role adaptation problems. Recent studies have successfully exploited the language understanding and reasoning capabilities of large language models (LLMs) to learn manipulation tasks. Impressed by the remarkable power of LLMs, this paper employs LLMs as role-specific reward designer for playing StarCraft II, making rewards more flexible and task-oriented. Firstly, we develop an interactive text and multi-agent RL environment to study real-time strategy generation in StarCraft II. Secondly, we use LLMs to interpret the game situation and understand agent roles from user instructions. Then, by assigning appropriate subtasks, LLMs quantify the completion of these subtasks to generate role-specific rewards. Further, credit assignment problem is addressed by introducing dynamic reward weights in value decomposition method. In StarCraft II maps, experiments show that role-aligned RL agents trained with our framework achieve superior policy performance, and win rate results demonstrates the effectiveness of our approach in decision-making for micromanagement and long-term planning. © 2025 IEEE.

关键词： Large Language Models Multi-agent Reinforcement Learning Reward design StarCraft II Value decomposition

来源：评论

学校读者我要写书评

暂无评论

AccelES: Accelerating Top-K SpMV for Embedding Similarity via Low-bit Pruning 31

AccelES: Accelerating Top-K SpMV for Embedding Similarity vi...

引用

31st IEEE International Symposium on High Performance computer Architecture, HPCA 2025

作者： Zhai, Jiaqi Shi, Xuanhua Huang, Kaiyi Ye, Chencheng Hu, Weifang He, Bingsheng Jin, Hai Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Wuhan430074 China National University of Singapore School of Computing 119077 Singapore

ISBN: (纸本)9798331506476

In the realm of recommendation systems, achieving real-time performance in embedding similarity tasks is often hindered by the limitations of traditional Top-K sparse matrix-vector multiplication (SpMV) methods, which suffer from high latency due to inefficient memory access patterns. This paper identifies these critical gaps and introduces AccelES, a novel approach that significantly enhances the efficiency of Top-K SpMV. Our method employs a two-stage calculation scheme: the first stage utilizes a compact, low-bit dataset to quickly identify the most relevant entries, while the second stage performs full-precision calculations solely on this pruned subset, thereby minimizing computational overhead. Furthermore, AccelES incorporates innovative matrix representations, Ultra-CSR and Random-CSR, which optimize memory bandwidth utilization. Experimental results demonstrate that AccelES accelerates performance, surpassing state-of-the-art FPGA, GPU, and CPU solutions by factors of 3.4×, 2.5×, and 153.3×, respectively, under controlled conditions. These advancements not only enhance processing speed but also significantly improve real-time performance in recommendation systems, establishing AccelES as a pivotal contribution to the field of Top-K sparse matrix-vector multiplication. © 2025 IEEE.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Terrain Scene Generation Using a Lightweight Vector Quantized Generative Adversarial Network

引用

IEEE Transactions on Big Data 2025年第3期11卷 988-1000页

作者： Wang, Yan Zhou, Huiyu Dong, Xinghui Ocean University of China School of Computer Science and Technology Qingdao266100 China University of Leicester School of Computing and Mathematical Sciences LeicesterLE1 7RH United Kingdom

Natural terrain scene images play important roles in the geographical research and application. However, it is challenging to collect a large set of terrain scene images. Recently, great progress has been made in image generation. Although impressive results can be achieved, the efficiency of the state-of-the-art methods, e.g., the Vector Quantized Generative Adversarial Network (VQGAN), is still dissatisfying. The VQGAN confronts two issues, i.e., high space complexity and heavy computational demand. To efficiently fulfill the terrain scene generation task, we first collect a Natural Terrain Scene Data Set (NTSD), which contains 36,672 images divided into 38 classes. Then we propose a Lightweight VQGAN (Lit-VQGAN), which uses the fewer parameters and has the lower computational complexity, compared with the VQGAN. A lightweight super-resolution network is further adopted, to speedily derive a high-resolution image from the image that the Lit-VQGAN generates. The Lit-VQGAN can be trained and tested on the NTSD. To our knowledge, either the NTSD or the Lit-VQGAN has not been exploited before.1 Experimental results show that the Lit-VQGAN is more efficient and effective than the VQGAN for the image generation task. These promising results should be due to the lightweight yet effective networks that we design. © 2015 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Based Trust Model for Inter-Domain Routing

引用

computers, Materials & Continua 2025年第3期82卷 4821-4839页

作者： Qiong Yang Li Ma Sami Ullah Shanshan Tu Hisham Alasmary Muhammad Waqas Faculty of Information Technology Beijing University of TechnologyBeijing100124China School of Information Science and Technology North China University of TechnologyBeijing100144China Department of Computer Science Shaheed Benazir Bhutto UniversitySheringal18050Pakistan Department of Computer Science King Khalid UniversityAbha61421Saudi Arabia School of Computing andMathematical Sciences Faculty of Engineering and ScienceUniversity of GreenwichLondonSE109LSUK School of Engineering Edith Cowan UniversityPerthWA 6027Australia

Border Gateway Protocol(BGP),as the standard inter-domain routing protocol,is a distance-vector dynamic routing protocol used for exchanging routing information between distributed Autonomous Systems(AS).BGP nodes,communicating in a distributed dynamic environment,face several security challenges,with trust being one of the most important issues in inter-domain *** research,which performs trust evaluation when exchanging routing information to suppress malicious routing behavior,cannot meet the scalability requirements of BGP *** this paper,we propose a blockchain-based trust model for inter-domain *** model achieves scalability by allowing the master node of an AS alliance to transmit the trust evaluation data of its member nodes to the *** BGP nodes can expedite the trust evaluation process by accessing a global view of other BGP nodes through the master node of their respective *** incorporate security service evaluation before direct evaluation and indirect recommendations to assess the security services that BGP nodes provide for themselves and prioritize to guarantee their security of routing *** forward the trust evaluation for neighbor discovery and prioritize the nodes with high trust as neighbor nodes to reduce the malicious exchange routing *** use simulation software to simulate a real BGP environments and employ a comparative experimental research approach to demonstrate the performance evaluation of our trust *** with the classical trust model,our trust model not only saves more storage overhead,but also provides higher security,especially reducing the impact of collusion attacks.

关键词： Inter-domain routing BGP security blockchain trust model trust mechanisms trust evaluation

来源：评论

学校读者我要写书评

暂无评论

Seer: Accelerating Block chain Transaction Execution by Fine-Grained Branch Prediction 51st

Seer: Accelerating Block chain Transaction Execution by Fine...

引用

51st International Conference on Very Large Data Bases, VLDB 2025

作者： Zhang, Shijie Cheng, Ru Liu, Xinpeng Xiao, Jiang Jin, Hai Li, Bo National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan430074 China Department of Computer Science and Engineering Hong Kong University of Science and Technology Hong Kong

Increasingly popular decentralized applications (dApps) with complex application logic incur significant overhead for executing smart contract transactions, which greatly limits public block chain performance. Pre-executing transactions off the critical path can mitigate substantial I/O and computation costs during execution. However, pre-execution does not yield any state transitions, rendering the system state inconsistent with actual execution. This inconsistency can lead to deviations in pre-execution paths when processing smart contracts with multiple state-related branches, thus diminishing pre-execution effectiveness. In this paper, we develop Seer, a novel public block chain execution engine that incorporates fine-grained branch prediction to fully exploit pre-execution effectiveness. Seerpredicts state-related branches using a two-level prediction approach, reducing inconsistent execution paths more efficiently than executing all possible branches. To enable effective reuse of pre execution results, Seer employs checkpoint-based fast-path execution, enhancing transaction execution for both successful and unsuccessful predictions. Evaluations with realistic block chain workloads demonstrate that Seer delivers an average of 27.7× transaction-level speedup and an overall 20.6× speedup in the execution phase over vanilla Ethereum, outperforming existing block chain execution acceleration solutions. © 2025, VLDB Endowment, All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

MeHyper: Accelerating Hypergraph Neural Networks by Exploring Implicit Dataflows 31

MeHyper: Accelerating Hypergraph Neural Networks by Explorin...

引用

31st IEEE International Symposium on High Performance computer Architecture, HPCA 2025

作者： Zhao, Wenju Yao, Pengcheng Chen, Dan Zheng, Long Liao, Xiaofei Wang, Qinggang Ma, Shaobo Li, Yu Liu, Haifeng Xiao, Wenjing Sun, Yufei Zhu, Bing Jin, Hai Xue, Jingling Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Wuhan430074 China National University of Singapore School of Computing 119077 Singapore Guangxi University School of Computer Electronics and Information NanNing530004 China University of New South Wales School of Computer Science and Engineering SydneyNSW2052 Australia

ISBN: (纸本)9798331506476

Hypergraph Neural Networks (HGNNs) are increasingly utilized to analyze complex inter-entity relationships. Traditional HGNN systems, based on a hyperedge-centric dataflow model, independently process aggregation tasks for hyperedges and vertices, leading to significant computational redundancy. This redundancy arises from recalculating shared information across different tasks. For the first time, we identify and harness implicit dataflows (i.e., dependencies) within HGNNs, introducing the microedge concept to effectively capture and reuse intricate shared information among aggregation tasks, thereby minimizing redundant computations. We have developed a new microedge-centric dataflow model that processes shared information as fine-grained microedge aggregation tasks. This dataflow model is supported by the Read-Process-Activate-Generate execution model, which aims to optimize parallelism among these tasks. Furthermore, our newly developed MeHyper, a microedge-centric HGNN accelerator, incorporates a decoupled pipeline for improved computational parallelism and a hierarchical feature management strategy to reduce off-chip memory accesses for large volumes of intermediate feature vectors generated. Our evaluation demonstrates that MeHyper substantially outperforms the leading CPUbased system PyG-CPU and the GPU-based system HyperGef, delivering performance improvements of 1,032.23 × and 10.51 ×, and energy efficiencies of 1,169.03 × and 9.96 ×, respectively. © 2025 IEEE.

关键词： Redundancy

来源：评论

学校读者我要写书评

暂无评论

Artificial intelligence algorithms for object detection and recognition in video and images

引用

Multimedia Tools and Applications 2025年 1-18页

作者： Dakshinamoorthy, Prabakar Rajaram, Gnanajeyaraman garg, Shruti Murugan, Prabhu Manimaran, A. Sundar, Ramesh Department of Data Science and Business System School of Computing SRM Institute of Science and Technology SRM Nagar Kattankulathur Chennai India Department of Computer Science Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Tamilnadu Chennai India Birla Institute of Technology Mesra Ranchi India Department of ECE Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Chennai602105 India Department of Computer Science Engineering Saveetha School Of EngineeringSaveetha Institute of Medical and Technical Sciences Chennai6021055 India Department of Netwoking and Communication School of Computing SRM Institute of Science and Technology SRM Nagar Kattankulathur Chennai India

The usage of machine learning and deep learning algorithms have necessitated Artificial Intelligence'. AI is aimed at automating things by limiting human interference. It is widely used in IT, healthcare, finance, and agriculture. It is achieved through several deep learning algorithms that reflect the human brain's intelligence. These AI algorithms can be manipulated according to changing needs and improved efficiency. This paper tries to utilize the developments made in AI technology to classify the images and recognize the objects present in them. One widely used AI algorithm is CNN (Convolutional Neural Networks). The CNN is a deep learning-based algorithm that consists of various layers that extract and filters the parameters present in the images. Some additional layers of ResNet50 and the CNN algorithm are used to extract the parameters to improve image recognition accuracy. The image dataset taken for training and testing the proposed model is imageNet. The images are initially processed before sending them to the proposed model. The proposed model is trained, validated, and tested through the images obtained after the initial processing. The same process is repeated several times until getting the maximum accuracy. The accuracy of the proposed model in terms of image recognition is recorded. The obtained results are compared with other image classification algorithms like VGG16 and VGG19. It is concluded that the proposed model outperforms other traditional methods in terms of accuracy. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

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：