CVPR 2016 Tutorial


Low-Rank and Sparse Modeling for Visual Analytics


Rene Vidal, Ehsan Elhamifar, Zhouchen Lin, Jiashi Feng, Sheng Li, Yun Fu


Low-rank and sparse modeling are emerging mathematical tools dealing with uncertainties of real-world visual data. Leveraging on the underlying structure of data, low-rank and sparse modeling approaches have achieved impressive performance in many visual learning tasks. In this tutorial, we will introduce in detail the latest subspace clustering techniques. For both sparsity based and low-rankness based approaches, we will present some representative subspace clustering models, analyze their theoretical properties, such as exact recovery and closed-form solutions, and present applications in various real problems in computer vision. We will also discuss convex and nonconvex algorithms and scalable methods that can effectively address the problem of sparse and low-rank recovery for visual data.

Date: June 26, 2016


08:30-08:45 Introduction to Subspace Clustering

08:45-10:00 Sparse Subspace Clustering

a)    Modeling for High-Dimensional Data Analysis

b)    Scalable Sparse Subspace Clustering

10:00-10:30 Coffee Break

10:30-12:00 Low Rank Subspace Clustering

a)   Representative Models

b)   Analysis on LRR

c)    Applications

d)   Block-Diagonal Structure in Subspace Clustering

12:00-02:00 Lunch Break

02:00-03:30 Algorithms & More Models

a)   Optimization and Analysis

b)   Scalable Algorithms

c)    More Models

03:30-04:00 Coffee Break

04:00-04:45 Applications & Conclusions

a)   Applications

b)   Conclusions