检索结果-内蒙古大学图书馆

Densely activated self-attention for semantic segmentation

Pattern Recognition 2025年

作者： Liwen Xiao Wenze Liu Zhicheng Wang Yiran Wang Zhiyu Pan Hao Lu Zhiguo Cao Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan 430074 China

In semantic segmentation, pixels often share the same mask labels across a vast region. However, in recent prevalent transformer-based models, predictions frequently suffer from incompleteness or discontinuities. The dilemma is caused by sparse activation of vanilla self-attention during feature extraction. During the vanilla self-attention, each query over-focuses on a small number of relevant keys but neglects numerous keys sharing the same category with it, restricting the capture of universal feature representation for its category. Such sparsely activated self-attention will further cause unignorable feature differences among tokens sharing the same class in feature maps, introducing noise on final predictions. To reduce such differences, we propose the Densely Activated self-attention Module (DAM), a novel pluggable module designed to generate densely activated self-attention. Inserted after the encoder, it encourages each query to attend to a broader range of keys, obtaining more consistent features. Experimental results on three widely used benchmarks with six different baselines demonstrate that DAM consistently improves performance with a negligible increase in parameters and FLOPs. Our work provides a new perspective on the behavior of self-attention in semantic segmentation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Course Homework Reform in Universities Based on Extended Reality 17th

Course Homework Reform in Universities Based on Extended Rea...

引用

17th International Conference on Computer Science and Education, ICCSE 2022

作者： Fu, Jinrong Liu, Yiwen Mi, Chunqiao Peng, Xiaoning Xiao, Jianhua School of Computer and Artificial Intelligence Huaihua University Hunan Huaihua418000 China Key Laboratory of Wuling-Mountain Health Big Data Intelligent Processing and Application in Hunan Province Universities Hunan Huaihua418000 China Key Laboratory of Intelligent Control Technology for Wuling-Mountain Ecological Agriculture in Hunan Province Hunan Huaihua418000 China

ISBN: (纸本)9789819924486

In view of the serious issues commonly existing with coursework in Chinese universities at present, such as its original function weakening or being suppressed, its form being too abstract and lack of elaborate design. In order to resolve these problems observed in usual homework, this paper proposes a reform scheme for cloud coursework based on Extended Reality and Bloom model. It mainly includes five key parts, such as interaction, scene, comprehensive evaluation, digital twin and data analysis. In addition, the reform’s effectiveness is illustrated through amassing and comparing the feedback of college students for teaching. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Augmented reality

来源：评论

学校读者我要写书评

暂无评论

Fast Intra Mode Prediction Algorithms for SCBS in VVC SCC

Fast Intra Mode Prediction Algorithms for SCBS in VVC SCC

引用

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Dayong Wang Yishen Deng Weisheng Li Xin Lu Frederic Dufaux Bo Hang Ce Zhu Key Laboratory of Big Data Intelligent Computing Chongqing University of Posts and Telecommunications Guangxi Key Laboratory of Machine Vision and Intelligent Control Wuzhou University Chongqing Key Laboratory of Image Cognition Chongqing University of Posts and Telecommunications Faculty of Computing Engineering and Media (CEM) De Montfort University Université Paris-Saclay CNRS CentraleSupélec Laboratoire des Signaux et systèmes The Computer School Hubei University of Arts and Science School of Information and Communication Engineering University of Electronic Science and Technology of China

Versatile Video Coding (VVC) now supports Screen Content Coding (SCC) by integrating two efficient coding modes: Intra Block Copy (IBC) and Palette (PLT). However, the numerous modes and the Quad-Tree Plus Multi-Type Tree (QTMT) structure inherent to VVC contribute to a very high coding complexity. To effectively reduce the computational complexity of VVC SCC, we propose a fast Intra mode prediction algorithm for VVC SCC. More specifically, we first use the difference of minimum Sum of Absolute Transformed Differences (SATD) value of four Directional Modes (DMs) of Intra and the SATD value of the IBC-merge mode to determine whether to early skip Intra checking. Subsequently, we use a decision tree to determine whether to early terminate the checking after block differential pulse coded modulation (BDPCM). Finally, we employ a decision tree to determine whether to early skip multiple transform selection (MTS) and low frequency non-separable transform (LFNST) checking. The results demonstrate that our algorithm achieves an average encoding time reduction of 34.34% with a negligible Bjøntegaard delta bitrate increase of 0.46%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Teaching Reform and Practice for the Course of Web Application Development Technology Under the Background of Online Education 17th

Teaching Reform and Practice for the Course of Web Applicati...

引用

17th International Conference on Computer Science and Education, ICCSE 2022

作者： Dong, Yi Xiaoning, Peng Xian, Zhang Chunqiao, Mi Yiwen, Liu Taiguo, Qu School of Computer and Artificial Intelligence Huaihua University Hunan Huaihua418000 China Key Laboratory of Wuling-Mountain Health Big Data Intelligent Processing and Application in Hunan Province Universities Hunan Huaihua418000 China Key Laboratory of Intelligent Control Technology for Wuling-Mountain Ecological Agriculture in Hunan Province Hunan Huaihua418000 China

ISBN: (纸本)9789819924455

With the advent of knowledge economy, online education relies on the Internet and mobile terminals, it breaks the limitation of time and space of traditional education. Online education has gradually becoming a new way for people to learn. Under the special background of COVID-19 since 2020, online education has been well applied. The course of Web Application Development Technology is not only a course which is the core course in software development for computer majors, but also is a course with strong practicality and application. This paper analyzes the problems existing in the online teaching of the courses of web course under the background of online education, and presents a new scheme of the teaching reform measures for the course according to the characteristics of online education and the course of Web Application Development Technology which is well applied by the author’s own practices. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Graph Convolutional Networks based Muti-Label Deep Cross-Modal Hashing 13

Graph Convolutional Networks based Muti-Label Deep Cross-Mod...

引用

13th International Conference on Information Science and Technology, ICIST 2023

作者： Peng, Yi Zhang, Nian Jiang, Xin Xiong, Jiang School of Electronic and Information Engineering Chongqing Three Gorges University Chongqing40044 China Department of Electrical and Computer Engineering University of the District of Columbia WashingtonDC20008 United States Key Laboratory of Intelligent Information Processing and Control of Chongqing Municipal Institutions of Higher Education Chongqing Three Gorges University Chongqing40044 China College of Mathematics and Statistics Chongqing Three Gorges University Chongqing40044 China

ISBN: (纸本)9798350313925

Recently, multi-label deep cross-modal hashing (MDCH), which incorporates deep neural networks, hashing and multi-label learning for cross-modal retrieval tasks, has achieved excellent cross-modal retrieval results and thus became a highly popular area of research. Nevertheless, many existing MDCH methods concentrate on extracting information from multi-modal data, while neglecting the abundant semantic information in multiple labels. Few MDCH methods incorporate multi-label information, but they often treat labels as independent entities, ignoring the relationships between categories, which hinders the establishment of semantic connections among multi-modal data. In order to tackle the aforementioned challenges, we propose a graph convolutional networks based multi-label deep cross-modal hashing method (GMCH) in this paper. GMCH leverages two deep neural networks to generate hash representations from the original image-text pairs, during this process, a graph convolutional network is introduced to capture the category correlations of multi-labels and supervise the training of hash mapping. Experimental results on two commonly employed datasets validate the efficacy of the proposed GMCH method. You can find the code for our proposed GCMH at https://***/licher12/***. © 2023 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Vision-based Autonomous Landing of UAV on Dynamic Apron

Vision-based Autonomous Landing of UAV on Dynamic Apron

引用

IEEE International Conference on Unmanned Systems (ICUS)

作者： Pengyu Yue Jing Xin Zhijie Mao Yongchang Lu Youmin Zhang Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing Xi'an University of Technology Xi'an P.R.China Department of Mechanical Industrial & Aerospace Engineering Concordia University Montreal QC Canada

Aiming at the landing problem of Unmanned Aerial Vehicle (UAV) without landmarks and global coordinates, this paper proposes an efficient vision-based UAV autonomous landing solution. Firstly, image information is obtained from the bottom visual sensor of the UAV, and the YOLOV5 object detection method is employed to determine the initial position of the dynamic apron. Subsequently, an improved dynamic apron tracking method based on Spatio-Temporal Transformer is designed. This method updates an initial template through target detection algorithm and utilizes multi-source perceptual data for confidence evaluation and dynamic template updates. It outputs the distribution estimation of the predicted bounding box corner coordinates, and achieves the high reliability continuous positional smoothing estimation and dynamic apron loss re-detection based on UAV vision. Finally, the UAV calculates the horizontal control quantity based on the horizontal relative position between the UAV and the apron obtained. The vertical control quantity is calculated based on the vertical relative position between the UAV and the apron. The vertical relative position is estimated using laser ranging and visual odometry techniques. Based on the results of horizontal tracking control and vertical regulation, velocity commands for the UAV in all three directions are generated, enabling the UAV to achieve autonomous landing on the dynamic apron. This paper achieves precise landing for dynamic apron in real-world environments. The experimental results demonstrate that this method avoids the limitations of labeled data and global information on autonomous landing of UAV. Even under GPS-denied conditions, the UAV can generate control commands based on visual data and achieve autonomous landing for randomly moving apron.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Implicit Ray-Transformers for Multi-view Remote Sensing image Segmentation

arXiv

引用

arXiv 2023年

作者： Qi, Zipeng Chen, Hao Liu, Chenyang Shi, Zhenwei Zou, Zhengxia The Image Processing Center School of Astronautics The Beijing Key Laboratory of Digital Media The State Key Laboratory of Virtual Reality Technology and Systems Beihang University Beijing100191 China The Shanghai Artificial Intelligence Laboratory Shanghai200232 China The Department of Guidance Navigation and Control School of Astronautics Beihang University Beijing100191 China

The mainstream CNN-based remote sensing (RS) image semantic segmentation approaches typically rely on massive labeled training data. Such a paradigm struggles with the problem of RS multi-view scene segmentation with limited labeled views due to the lack of considering 3D information within the scene. In this paper, we propose "Implicit Ray-Transformer (IRT)" based on Implicit Neural Representation (INR), for RS scene semantic segmentation with sparse labels (such as 4-6 labels per 100 images). We explore a new way of introducing multi-view 3D structure priors to the task for accurate and view-consistent semantic segmentation. The proposed method includes a two-stage learning process. In the first stage, we optimize a neural field to encode the color and 3D structure of the remote sensing scene based on multi-view images. In the second stage, we design a Ray Transformer to leverage the relations between the neural field 3D features and 2D texture features for learning better semantic representations. Different from previous methods that only consider 3D prior or 2D features, we incorporate additional 2D texture information and 3D prior by broadcasting CNN features to different point features along the sampled ray. To verify the effectiveness of the proposed method, we construct a challenging dataset containing six synthetic sub-datasets collected from the Carla platform and three real sub-datasets from Google Maps. Experiments show that the proposed method outperforms the CNN-based methods and the state-of-the-art INR-based segmentation methods in quantitative and qualitative metrics. Ablation study shows that under limited label conditions, the combination of the 3D structure prior and 2D texture can significantly improve the performance and effectively complete missing semantic information in novel views. Experiments also demonstrate the proposed method could yield geometry-consistent segmentation results against illumination changes and viewpoint changes. Our dat

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Time Series Anomaly Detection Based on Adversarial Interpolation and Pseudo-anomaly Calibration

Unsupervised Time Series Anomaly Detection Based on Adversar...

引用

International Conference on control, Automation and Robotics (ICCAR)

作者： Xinwei Chen Xiaohui Lin Zuoyong Li Haoyi Fan Fujian Provincial Key Laboratory of Information Processing and Intelligent Control College of Computer and Control Engineering Minjiang University Fuzhou China College of Computer and Data Science Fuzhou University Fuzhou China School of Computer and Artificial Intelligence Zhengzhou University Zhengzhou China

Existing work on unsupervised time series anomaly detection relies on clean training datasets (free of anomalies), which is often violated in practice. In this work, we propose an unsupervised time series anomaly detection method based on adversarial interpolation and pseudo-anomaly calibration for training an anomaly detector under data con-tamination. Specifically, the normality description learned from the model is improved by using the following two proposed calibration methods: (1) by enforcing adversarial interpolation to learn an effective normality description from the typical samples rather than the edge or abnormal samples, and (2) by discriminating the original samples from the simulated pseudo-anomaly samples to learn the exact normality description boundary. These two calibration ways result in the normality description that prevents data contamination. Experimental results compared with baselines on public time series datasets demonstrate the effectiveness of our proposed method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Autonomous Exploration and Mapping for Mobile Robots via Cumulative Curriculum Reinforcement Learning

Autonomous Exploration and Mapping for Mobile Robots via Cum...

引用

IEEE/RSJ International Conference on intelligent Robots and Systems (IROS)

作者： Zhi Li Jinghao Xin Ning Li Department of Automation Shanghai Jiao Tong University Shanghai P.R. China Ministry of Education of China Key Laboratory of System Control and Information Processing Shanghai China Shanghai Engi-neering Research Center of Intelligent Control and Management Shanghai China

Deep reinforcement learning (DRL) has been widely applied in autonomous exploration and mapping tasks, but often struggles with the challenges of sampling efficiency, poor adaptability to unknown map sizes, and slow simulation speed. To speed up convergence, we combine curriculum learning (CL) with DRL, and first propose a Cumulative Curriculum Reinforcement Learning (CCRL) training framework to alleviate the issue of catastrophic forgetting faced by general CL. Besides, we present a novel state representation, which considers a local egocentric map and a global exploration map resized to the fixed dimension, so as to flexibly adapt to environments with various sizes and shapes. Additionally, for facilitating the fast training of DRL models, we develop a lightweight grid-based simulator, which can substantially accelerate simulation compared to popular robot simulation platforms such as Gazebo. Based on the customized simulator, comprehensive experiments have been conducted, and the results show that the CCRL framework not only mitigates the catastrophic forgetting problem, but also improves the sample efficiency and generalization of DRL models, compared to general CL as well as without a curriculum. Our code is available at https://***/BeamanLi/CCRL_Exploration.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Channel and space-based joint rate allocation algorithm

Channel and space-based joint rate allocation algorithm

引用

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Dayong Wang Chao Yuan Yu Sun Xin Lu Hui Guo Frederic Dufaux Ce Zhu Key Laboratory of Big Data Intelligent Computing Chongqing University of Posts and Telecommunications Guangxi Key Laboratory of Machine Vision and Intelligent Control Wuzhou University Chongqing Key Laboratory of Image Cognition Chongqing University of Posts and Telecommunications China Department of Computer Science University of Central Arkansas Faculty of Computing Engineering and Media (CEM) De Montfort University UK Université Paris-Saclay CNRS CentraleSupélec Laboratoire Des Signaux et Systèmes France School of Information and Communication Engineering University of Electronic Science and Technology of China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Rate control is a critical component for image and video compression Particularly under limited network bandwidth conditions, bitrate control is essential to ensure efficient image transmission by effectively allocation channel resources. In this research, since both Channel and Spatial have relationship with rate allocation, we first propose a joint Channel-wise and Spatial-wise Quantization scheme to determine optimal quantization parameters. Subsequently, we develop a quantization step estimation network to obtain parameters to efficiently allocate rate according to target rate. Experiments demonstrate that our algorithm significantly improve compressed image quality with minimal bitrate distortion and achieve accurate rate control with nearly 3% average bitrate error.

关键词： image quality Quantization (signal) image coding Bit rate Signal processing algorithms Estimation Rate-distortion Video compression Resource management Speech processing

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：