ACCV 2014 Tutorial
Advanced Sparse Representation Models for Image and Video Analysis
Speaker: Shenghua Gao, Assistant professor, ShanghaiTech University, China
Kui Jia, Research Scientist, Advanced Digital Sciences Center (ADSC), Singapore
Tianzhu Zhang, Assistant professor, Institute of Automation, Chinese Academy of Sciences
Weisheng Dong, Associate professor, Xidian University, China
Date: Nov. 2, 2014
Duration: Half day
Intrinsic structures of high-dimensional visual data often have the properties of low dimensionality, sparsity, or degeneracy. By discovering and properly harnessing these intrinsic structures, groundbreaking results have been achieved in the past decade on diverse applications in the domains of signal/image processing, computational neuroscience, computer vision, and machine learning. This tutorial aims to present the very recent breakthroughs in computer vision and image processing research. These results are achieved by leveraging new powerful, more advanced sparse representation models and by developing efficient large-scale optimization algorithms.
Session 1: Theory and algorithms of sparse coding and its extensions
This session will introduce the basic theory of sparse/low-rank recovery, with its generalization to more advanced sparse models which considers other structures of the data, for example, locality information, group sparse structure, nonlinear structure, tree-guided structure, etc. We introduce these concepts in the context of signal processing and from the perspectives of geometric intuition, formal definition, and recovery conditions, with motivating applications.
Session 2: Advanced sparse representation models for image processing
This session will present the very recent breakthroughs in image processing. These results are essentially obtained by properly harnessing these rich low-dimensional structures prevailing in natural images, using carefully designed more advanced models with sparsity/low-rank constriants, for various low-level vision tasks including image de-noising, de-blurring, super-resolution, and compressive sensing, etc. We will specially show that how local sparsity/low-rank models and non-local self-similarity of natural images are exploited to achieve these breakthrough results.
Session 3: Advanced sparse representation models for object/face recognition
This session will present striking results recently obtained in computer vision research. We cover a variety of mainstream vision applications ranging from face/object recognition, object alignment, feature correspondence/matching, tracking, to unsupervised object discovery and ambiguous learning. We will introduce the respective nature of these problems and explain the principles in designing more advanced sparse representation models in order to harness their problem nature and achieve striking performance.
Session 4: Advanced sparse representation models for object tracking
This session will also discuss how the advanced sparse representation models can be used for video tracking tasks.
Shenghua Gao is an assistant professor in ShanghaiTech University, China. He received the B.E. degree from the University of Science and Technology of China in 2008 (outstanding graduates), and received the Ph.D. degree from the Nanyang Technological University in 2012. From Jun 2012 to Aug 2014, he worked as a research scientist in Advanced Digital Sciences Center, Singapore. His research interests include computer vision and machine learning. He has published more than 20 papers on object and face recognition related topics in many international conferences and journals, including IEEE T-PAMI, IJCV, IEEE TIP, IEEE TNNLS, IEEE TMM, IEEE TCSVT, CVPR, ECCV, etc.). He was awarded the Microsoft Research Fellowship in 2010.
Kui Jia received the B.Eng. degree in marine engineering from Northwestern Polytechnic University, China, in 2001, the M.Eng. degree in electrical and computer engineering from National University of Singapore in 2003, and the Ph.D. degree in computer science from Queen Mary, University of London, London, U.K., in 2007. Since October 2011, he has been working as a Research Scientist at Advanced Digital Sciences Center (ADSC), Singapore, which is a joint research center between the University of Illinois at Urbana-Champaign (UIUC) and the Agency for Science, Technology and Research (A-STAR), Singapore. He also holds a joint appointment with the Coordinated Science Laboratory of UIUC. Before joining ADSC, he was with The Chinese University of Hong Kong, and Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences. His research interests are in computer vision, machine learning, and image processing.
Tianzhu Zhang received the B.Eng. degree in communications and information technology from Beijing Institute of Technology, Beijing, China, in 2006, and the Ph.D. degree in pattern recognition and intelligent systems from the Institute of Automation, Chinese Academy of Sciences in 2011. After graduation, he worked as a research postdoctoral fellow at Advanced Digital Sciences Center (ADSC), Singapore, which is a joint research center between the University of Illinois at Urbana-Champaign (UIUC) and the Agency for Science, Technology and Research (A-STAR), Singapore. Since October 2013, he has been working as a Research Scientist. His current research interests include computer vision and multimedia, especially action recognition, object classification and object tracking.
Weisheng Dong received the B. S degree in communication engineering from Huazhong University of science and technology, Wuhan, China in 2004 and the Ph. D. degree in circuit and systems from Xidian University, Xi’an, China in 2010. From Jan. 2009 to Jun. 2010, he was a research assistant / associate with Dep. of Computing, the Hong Kong Polytechnic University, Hong Kong. From Aug. 2012 to Feb. 2013 he was a visiting researcher with Microsoft Research Asia. In Sep. 2010 he joined the school of electronic engineering, Xidian University as a Lecturer, and has been an associate professor since Jun. 2012. His research interests include inverse problems in image processing, sparse representation, and image compression. He was the recipient of the best paper award at SPIE VCIP 2010.
 Weisheng Dong, Guangming Shi, Xiaocheng Hu, and Yi Ma, "Nonlocal sparse and low-rank regularization for optical flow estimation," IEEE Trans. on Image Processing (TIP), in press, 2014.
 Weisheng Dong, Lei Zhang, Rastislav Lukac, and Guangming Shi, “Sparse representation based image interpolation with nonlocal autoregressive modeling,” IEEE Trans. on Image Processing (TIP), vol. 22, no. 4, pp. 1382-1394, Apr. 2013.
 Weisheng Dong, Guangming Shi, Xin Li, Yi Ma, and Feng Huang, "Comressive sensing via nonlocal low-rank regularization", IEEE Trans. on Image Processing (TIP), vol. 23, no. 8, pp. 3618-3612, Aug. 2014.
 Weisheng Dong, Lei Zhang, Guangming Shi, and Xin Li, “Nonlocal centralized sparse representation for image restoration,” IEEE Trans. on Image Processing (TIP), vol. 22, no. 4, pp. 1620-1630, Apr. 2013.
 Weisheng Dong, Lei Zhang, Guangming Shi, and Xiaolin Wu, “Image deblurring and super-resolution by adaptive sparse domain selection and adaptive regularization,” IEEE Trans. On Image Processing (TIP), vol. 20, no. 7, pp. 1838-1857, July 2011.
 Kui Jia, Xiaogang Wang, and Xiaoou Tang, "Image Transformation based on Learning Dictionaries across Image Spaces", IEEE Trans. on Pattern Analysis and Machine Intelligence (TPAMI), Vol. 35, Issue 2, pp. 367-380, 2013.
 Kui Jia, Tsung-Han Chan, and Yi Ma, "Robust and Practical Face Recognition via Structured Sparsity", in Proc. European Conference on Computer Vision (ECCV), 2012.
 Kui Jia, Tsung-Han Chan, Zinan Zeng, and Yi Ma, "ROML: A Robust Feature Correspondence Approach for Matching Objects in A Set of Images", arXiv:1403.7877.
 Zinan Zeng, Shijie Xiao, Kui Jia, Tsung-Han Chan, Shenghua Gao, Dong Xu, and Yi Ma, "Learning by Associating Ambiguously Labeled Images", in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2013.
 Tianzhu Zhang, Bernard Ghanem, Si Liu, Narendra Ahuja."Robust Visual Tracking via Structured Multi- Task Sparse Learning," International Journal of Computer Vision (IJCV), Volume: 101, Issue: 2, pp. 367- 383, 2013.
 Tianzhu Zhang, Bernard Ghanem, Si Liu, Narendra Ahuja."Robust Visual Tracking via Multi-Task Sparse Learning", IEEE International Conference on Computer Vision and Pattern Recognition (CVPR Oral 2.5%), 2012.
 Shenghua Gao, IvorWai-Hung Tsung, Yi Ma. “Learning Category-Specific Dictionary and Shared Dictionary for Fine-Grained Image Categorization”, IEEE Transactions on Image Processing (TIP),23(2):623 - 634, Feb 2014.
 Shenghua Gao, IvorWai-Hung Tsang and Liang-Tien Chia.” Sparse Representation with Kernels”, IEEE Transactions on Image Processing (TIP). 22(2):423 - 434, Feb 2013.
 Shenghua Gao, Ivor Wai-Hung Tsang and Liang-Tien Chia. “Laplacian Sparse Coding, Hypergraph Laplacian Sparse Coding, and Applications”, IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 35:(1):92-104, Jan 2013.
 Zhixiang Ren, Shenghua Gao, Liang-Tien Chia, Deepu Rajan and Yun Huang. “Regularized Feature Reconstruction for Spatio-temporal Saliency Detection”, IEEE Transactions on Image Processing (TIP).22(8):3120 - 3132, August 2013.
 Shenghua Gao, Kui Jia, Liansheng Zhuang, Yi Ma, “Neither Global Nor Local: Regularized Patch-based Representation for Single Sample Per person Face Recognition”, To appear in International Journal of Computer Vision (IJCV).
 Tianzhu Zhang, Si Liu, Narendra Ahuja, Ming-Hsuan Yang, Bernard Ghanem."Robust Visual Tracking via Consistent Low-Rank Sparse Learning", International Journal of Computer Vision (IJCV), 2014