检索结果-内蒙古大学图书馆

Fast and accurate single-image depth estimation on mobile devices, mobile AI 2021 challenge: Report

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ignatov, Andrey Malivenko, Grigory Plowman, David Shukla, Samarth Timofte, Radu Zhang, Ziyu Wang, Yicheng Huang, Zilong Luo, Guozhong Yu, Gang Fu, Bin Wang, Yiran Li, Xingyi Shi, Min Xian, Ke Cao, Zhiguo Du, Jin-Hua Wu, Pei-Lin Ge, Chao Yao, Jiaoyang Tu, Fangwen Li, Bo Yoo, Jung Eun Seo, Kwanggyoon Xu, Jialei Li, Zhenyu Liu, Xianming Jiang, Junjun Chen, Wei-Chi Joya, Shayan Fan, Huanhuan Kang, Zhaobing Li, Ang Feng, Tianpeng Liu, Yang Sheng, Chuannan Yin, Jian Benavides, Fausto T. Computer Vision Lab ETH Zurich Switzerland Ltd AI Witchlabs Switzerland Tencent GY-Lab China Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology China Nanjing Artificial Intelligence Chip Research Institute of Automation Chinese Academy of Sciences China Black Sesame Technologies Inc. Singapore Singapore Visual Media Lab KAIST Korea Republic of Harbin Institute of Technology China Peng Cheng Laboratory China Multimedia and Computer Vision Laboratory National Cheng Kung University Taiwan Samsung Research UK United Kingdom OPPO Research Institute China ETH Zurich Switzerland

Depth estimation is an important computer vision problem with many practical applications to mobile devices. While many solutions have been proposed for this task, they are usually very computationally expensive and thus are not applicable for on-device inference. To address this problem, we introduce the first Mobile AI challenge, where the target is to develop an end-to-end deep learning-based depth estimation solutions that can demonstrate a nearly real-time performance on smartphones and IoT platforms. For this, the participants were provided with a new large-scale dataset containing RGB-depth image pairs obtained with a dedicated stereo ZED camera producing high-resolution depth maps for objects located at up to 50 meters. The runtime of all models was evaluated on the popular Raspberry Pi 4 platform with a mobile ARM-based Broadcom chipset. The proposed solutions can generate VGA resolution depth maps at up to 10 FPS on the Raspberry Pi 4 while achieving high fidelity results, and are compatible with any Android or Linux-based mobile devices. A detailed description of all models developed in the challenge is provided in this paper. Copyright © 2021, The Authors. All rights reserved.

关键词： computer operating systems

Multi-level graph convolutional network with automatic graph learning for hyperspectral image classification

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wan, Sheng Gong, Chen Pan, Shirui Yang, Jie Yang, Jian PCA Lab Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education Jiangsu Key Laboratory of Image Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing210094 China Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai200240 China Faculty of Information Technology Monash University ClaytonVIC3800 Australia

Nowadays, deep learning methods, especially the Graph Convolutional Network (GCN), have shown impressive performance in hyperspectral image (HSI) classification. However, the current GCN-based methods treat graph construction and image classification as two separate tasks, which often results in suboptimal performance. Another defect of these methods is that they mainly focus on modeling the local pairwise importance between graph nodes while lack the capability to capture the global contextual information of HSI. In this paper, we propose a Multi-level GCN with Automatic Graph Learning method (MGCN-AGL) for HSI classification, which can automatically learn the graph information at both local and global levels. By employing attention mechanism to characterize the importance among spatially neighboring regions, the most relevant information can be adaptively incorporated to make decisions, which helps encode the spatial context to form the graph information at local level. Moreover, we utilize multiple pathways for local-level graph convolution, in order to leverage the merits from the diverse spatial context of HSI and to enhance the expressive power of the generated representations. To reconstruct the global contextual relations, our MGCN-AGL encodes the long range dependencies among image regions based on the expressive representations that have been produced at local level. Then inference can be performed along the reconstructed graph edges connecting faraway regions. Finally, the multi-level information is adaptively fused to generate the network output. In this means, the graph learning and image classification can be integrated into a unified framework and benefit each other. Extensive experiments have been conducted on three real-world hyperspectral datasets, which are shown to outperform the state-of-the-art methods. Copyright © 2020, The Authors. All rights reserved.

关键词： image classification

Technique Report of CVPR 2024 PBDL Challenges

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lu, Xiaoqiang Jiao, Licheng Liu, Fang Liu, Xu Li, Lingling Ma, Wenping Yang, Shuyuan Xie, Haiyang Zhao, Jian Huang, Shihua Cheng, Peng Shen, Xi Wang, Zheng An, Shuai Zhu, Caizhi Li, Xuelong Chen, Linwei Fu, Ying Zhang, Tao Liu, Yu Yan, Chenggang Wang, Zichun Wang, Qinliang Liu, Yang Yu, Xinyue Jia, Sen Zhang, Junpei Chen, Puhua Chen, Xiang Li, Hao Pan, Jinshan Xie, Chuanlong Chen, Hongming Li, Mingrui Deng, Tianchen Huang, Jingwei Li, Yufeng Wan, Fei Xu, Bingxin Cheng, Jian Liu, Hongzhe Xu, Cheng Zou, Yuxiang Pan, Weiguo Dai, Songyin Jiang, Linyan Song, Bingyi An, Zhuoyu Lei, Haibo Luo, Qing Song, Jie Liu, Yuan Li, Qihang Zhang, Haoyuan Wang, Lingfeng Chen, Wei Luo, Aling Li, Cheng Cao, Jun Chen, Shu Dou, Zifei Liu, Xinyu Zhang, Jing Zhang, Kexin Yang, Yuting Zhang, Liwen Xu, Zhe Gou, Dingyong Li, Cong Xu, Senyan Zhang, Yunkang Jiang, Siyuan Liu, Qinglin Yu, Wei Lv, Xiaoqian Li, Jianing Zhang, Shengping Ji, Xiangyang Zou, Yunhao Chen, Yuanpei Zhang, Yuhan Peng, Weihang Gou, Xuejian Zhao, Shizhan Zhang, Yanzhao Yan, Libo Guo, Yuwei Li, Guoxin Gao, Qiong Che, Chenyue Sun, Long School of Artificial Intelligence Xidian University China School of Computer Science Wuhan University China Intellindust China Beijing Forestry University China China Telecom China Harbin Institute of Technology China Northwestern Polytechnical University China China Telecom China Beijing Institute of Technology China Lishui Institute of Hangzhou Dianzi University China Tsinghua University China Intelligent Perception and Image Understanding Lab Xidian University China Nanjing University of Science and Technology China Shenyang Aerospace University China Dalian University of Technology China Shanghai Jiao Tong University China University of Electronic Science and Technology of China China Beijing Key Laboratory of Information Service Engineering Beijing Union University Beijing100101 China College of Robotics Beijing Union University China Tencent China Sanechips Technology Co. LTD. China Xiaomi Inc. China ZTE Corporation China University of Science and Technology of China China Peking University China Intelligent Science & Technology Academy of CASIC China

The intersection of physics-based vision and deep learning presents an exciting frontier for advancing computer vision technologies. By leveraging the principles of physics to inform and enhance deep learning models, we can develop more robust and accurate vision systems. Physics-based vision aims to invert the processes to recover scene properties such as shape, reflectance, light distribution, and medium properties from images. In recent years, deep learning has shown promising improvements for various vision tasks, and when combined with physics-based vision, these approaches can enhance the robustness and accuracy of vision systems. This technical report summarizes the outcomes of the Physics-Based vision Meets Deep Learning (PBDL) 2024 challenge, held in CVPR 2024 workshop. The challenge consisted of eight tracks, focusing on Low-Light Enhancement and Detection as well as High Dynamic Range (HDR) Imaging. This report details the objectives, methodologies, and results of each track, highlighting the top-performing solutions and their innovative approaches. © 2024, CC BY.

关键词： Machine vision

Nearest-neighbour Joint Probabilistic Data Association Filter Based on Random Finite Set

学校读者我要写书评

暂无评论

Nearest-neighbour Joint Probabilistic Data Association Filte...

International Conference on Control, Automation and Information Sciences (ICCAIS)

作者： Shuang Liang Yun Zhu Hao Li Maoguo Gong Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education Xidian University Xi'an China Key Laboratory of Modern Teaching Technology School of Computer Science Shaanxi Normal University Xi'an China

ISBN: (数字)9781728123110

ISBN: (纸本)9781728123127

The joint probabilistic data association (JPDA) filter is effective for multitarget, but it suffers from the track coalescence problem. To solve this problem, an improved nearest-neighbour JPDA filter based on random finite set is proposed. First, the standard JPDA filter is utilized to compute the target posterior density. Then, the posterior density is optimized by reordering the target index in each global association event using a novel nearest-neighbour method. Finally, the marginalized posterior densities of targets are obtained as independent Gaussian densities. Compared to conventional data association methods, the proposed approach needs less computing time and achieves satisfactory tracking accuracy.

关键词： Target tracking Radio frequency Clutter Filtering theory Covariance matrices Automation Probabilistic logic

Distributed Multi-Objective Community Detection in Large-Scale and Complex Networks

学校读者我要写书评

暂无评论

Distributed Multi-Objective Community Detection in Large-Sca...

International Conference on Computational Intelligence and Security

作者： Shuang Liang Hao Li Maoguo Gong Yue Wu Yun Zhu Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education Xidian University Xi'an China Key Laboratory of Modern Teaching Technology of Ministry of Education School of Computer Science Shaanxi Normal University Xi'an China

ISBN: (数字)9781728160924

ISBN: (纸本)9781728160931

Community detection has arisen as an important topic of many different research areas such as sociology, biology and computer science. However, with the appearance of "big data" and the rapid increasing size of real-world networks, a traditional evolutionary algorithm is unable or inefficient to solve the community detection problem in large-scale networks. In this paper, a distributed multi-objective evolutionary algorithm (DMOCD) for community detection is proposed. DMOCD is implemented based on Apache Spark and Resilient Distributed Datasets. The proposed distributed framework maintains a set of evolving populations (sub-populations) which evolve separately with different crossover and mutation parameters and an external repository as an elite archive to store the non-dominated individuals. A label propagation-based initialization method, a segmented crossover and mutation tactic for large-scale network are introduced. Experiments on both artificial and real-world networks prove that the proposed method is effective for community detection problems in small-scale networks and is able to process the large-scale networks which the stand-alone multi-objective evolutionary algorithms cannot deal with.

关键词： Sociology Statistics Optimization Partitioning algorithms Evolutionary computation Collaboration Complex networks

image Super-resolution Reconstruction Based on Bionic Intelligence

学校读者我要写书评

暂无评论

Image Super-resolution Reconstruction Based on Bionic Intell...

International Conference on image, vision and Computing (ICIVC)

作者： Xin Wang Qiong Wang Guofang Lv Jiangsu Key Laboratory of Image and Video Understanding for Social Safety Nanjing University of Science and Technology Nanjing China College of Computer and Information Hohai University Nanjing China

ISBN: (纸本)9781728123264

Traditional super-resolution reconstruction methods usually process a low resolution image as a whole, which ignores the differences among image patches, resulting in ineffective reconstruction result. In this paper, we adopt the concept of patch as the basic unit of super resolution reconstruction, and establish a novel image super-resolution reconstruction method based on bionic intelligence. The original low resolution image is initially reconstructed by bicubic interpolation. Then, a biological visual attention mechanism inspired approach is applied to the initial reconstruction result to detect the salient regions. Subsequently, a super-resolution convolutional neural network model is applied to the salient image patches so as to get the precise reconstruction results. Extensive experiments manifest that the proposed method outperforms many current state-of-the-art schemes in reconstruction quality, and achieves fast speed for practical application.

关键词： Super-resolution Bionic intelligence Saliency detection Convolutional neural network super-resolution Reconstructive Surgical Procedures image patch bionics Reconstruction algorithm Low-resolution image image resolution Reconstructions Intelligence

Efficient relaxed gradient support pursuit for sparsity constrained non-convex optimization

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Shang, Fanhua Wei, Bingkun Liu, Hongying Liu, Yuanyuan Zhuo, Jiacheng Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education School of Artificial Intelligence Xidian University China Department of Computer Science University of Texas Austin United States

Large-scale non-convex sparsity-constrained problems have recently gained extensive attention. Most existing deterministic optimization methods (e.g., GraSP) are not suitable for large-scale and high-dimensional problems, and thus stochastic optimization methods with hard thresholding (e.g., SVRGHT) become more attractive. Inspired by GraSP, this paper proposes a new general relaxed gradient support pursuit (RGraSP) framework, in which the sub-algorithm only requires to satisfy a slack descent condition. We also design two specific semi-stochastic gradient hard thresholding algorithms. In particular, our algorithms have much less hard thresholding operations than SVRGHT, and their average per-iteration cost is much lower (i.e., O(d) vs. O(dlog(d)) for SVRGHT), which leads to faster convergence. Our experimental results on both synthetic and real-world datasets show that our algorithms are superior to the state-of-the-art gradient hard thresholding methods. Copyright © 2019, The Authors. All rights reserved.

关键词： Constrained optimization

SoccerNet 2023 Challenges Results

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Cioppa, Anthony Giancola, Silvio Somers, Vladimir Magera, Floriane Zhou, Xin Mkhallati, Hassan Deliège, Adrien Held, Jan Hinojosa, Carlos Mansourian, Amir M. Miralles, Pierre Barnich, Olivier De Vleeschouwer, Christophe Alahi, Alexandre Ghanem, Bernard Van Droogenbroeck, Marc Kamal, Abdullah Maglo, Adrien Clapés, Albert Abdelaziz, Amr Xarles, Artur Orcesi, Astrid Scott, Atom Liu, Bin Lim, Byoungkwon Chen, Chen Deuser, Fabian Yan, Feng Yu, Fufu Shitrit, Gal Wang, Guanshuo Choi, Gyusik Kim, Hankyul Guo, Hao Fahrudin, Hasby Koguchi, Hidenari Ardö, Håkan Salah, Ibrahim Yerushalmy, Ido Muhammad, Iftikar Uchida, Ikuma Be'ery, Ishay Rabarisoa, Jaonary Lee, Jeongae Fu, Jiajun Yin, Jianqin Xu, Jinghang Nang, Jongho Denize, Julien Li, Junjie Zhang, Junpei Kim, Juntae Synowiec, Kamil Kobayashi, Kenji Zhang, Kexin Habel, Konrad Nakajima, Kota Jiao, Licheng Ma, Lin Wang, Lizhi Wang, Luping Li, Menglong Zhou, Mengying Nasr, Mohamed Abdelwahed, Mohamed Liashuha, Mykola Falaleev, Nikolay Oswald, Norbert Jia, Qiong Pham, Quoc-Cuong Song, Ran Hérault, Romain Peng, Rui Chen, Ruilong Liu, Ruixuan Baikulov, Ruslan Fukushima, Ryuto Escalera, Sergio Lee, Seungcheon Chen, Shimin Ding, Shouhong Someya, Taiga Moeslund, Thomas B. Li, Tianjiao Shen, Wei Zhang, Wei Li, Wei Dai, Wei Luo, Weixin Zhao, Wending Zhang, Wenjie Yang, Xinquan Ma, Yanbiao Joo, Yeeun Zeng, Yingsen Gan, Yiyang Zhu, Yongqiang Zhong, Yujie Ruan, Zheng Li, Zhiheng Huang, Zhijian Meng, Ziyu Belgium Saudi Arabia Sportradar Norway UCLouvain Belgium EPFL Switzerland EVS Broadcast Equipment Belgium Baidu Research United States Belgium Sharif University of Technology Iran Footovision France Zewail City of Science Technology and Innovation Egypt Université Paris-Saclay CEA France Universitat de Barcelona Spain Computer Vision Center Spain Nagoya University Japan Research Center for Applied Mathematics and Machine Intelligence Zhejiang Lab China AIBrain United States OPPO Research Institute China Germany Meituan China Tencent Youtu Lab China Amazon Prime Video Sport United States Sogang University Korea Republic of The University of Tokyo Japan Spiideo Sweden University of Tsukuba Japan School of Artificial Intelligence Beijing University of Posts and Telecommunications China Normandie Univ INSA Rouen LITIS France Shanghai Jiao Tong University China Key Laboratory of Intelligent Perception and Image Understanding The Ministry of Education Xidian University China NASK - National Research Institute Poland Robo Space China Tongji University China Sportlight Technology United Kingdom School of Control Science and Engineering Shandong University China lRomul Russia Aalborg University Denmark Turing AI Cultures GmbH Germany Information Systems Technology and Design Singapore University of Technology and Design Singapore Sun Yat-sen University China

The SoccerNet 2023 challenges were the third annual video understanding challenges organized by the SoccerNet team. For this third edition, the challenges were composed of seven vision-based tasks split into three main themes. The first theme, broadcast video understanding, is composed of three high-level tasks related to describing events occurring in the video broadcasts: (1) action spotting, focusing on retrieving all timestamps related to global actions in soccer, (2) ball action spotting, focusing on retrieving all timestamps related to the soccer ball change of state, and (3) dense video captioning, focusing on describing the broadcast with natural language and anchored timestamps. The second theme, field understanding, relates to the single task of (4) camera calibration, focusing on retrieving the intrinsic and extrinsic camera parameters from images. The third and last theme, player understanding, is composed of three low-level tasks related to extracting information about the players: (5) re-identification, focusing on retrieving the same players across multiple views, (6) multiple object tracking, focusing on tracking players and the ball through unedited video streams, and (7) jersey number recognition, focusing on recognizing the jersey number of players from tracklets. Compared to the previous editions of the SoccerNet challenges, tasks (2-3-7) are novel, including new annotations and data, task (4) was enhanced with more data and annotations, and task (6) now focuses on end-to-end approaches. More information on the tasks, challenges, and leader-boards are available on https://***. Baselines and development kits can be found on https://***/SoccerNet. © 2023, CC BY.

关键词： computer vision

DUAL-CHANNEL CONVOLUTIONAL NEURAL NETWORK FOR POLARIMETRIC SAR imageS CLASSIFICATION

学校读者我要写书评

暂无评论

DUAL-CHANNEL CONVOLUTIONAL NEURAL NETWORK FOR POLARIMETRIC S...

IEEE International Geoscience and Remote Sensing Symposium

作者： Wenqiang Hua Shuang Wang Wen Xie Yanhe Guo Xiaomin Jin School of Computer Science and Technology Xi’an University of Posts and Telecommunications Xi’an China Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China Xidian University Xi'an China

This paper presents a new dual-channel convolutional neural network (Dc-CNN) for Polarimetric synthetic aperture radar (PolSAR) image classification when labeled samples are small. First, a neighborhood minimum spanning tree (MST) is used to enlarge the labeled sample set. Then, in order to obtain the abundant spatial information, a new dual-channel CNN is designed to PolSAR image to acquire different spatial features. This network model contains two parallel CNN structures, which can extract different features used two multiscale convolution structure. Experiments results show that compared with other methods, the proposed method shows a satisfactory classification result.

关键词： Feature extraction Deep learning Convolution Training Convolutional neural networks image classification Covariance matrices