Auto-Grouped Sparse Representation for Visual Analysis

Abstract

In this work, we investigate how to automatically uncover the underlying group structure of a feature vector such that each group characterizes certain object-specific patterns, e.g. , visual pattern or motion trajectories from one object. By mining the group structure, we can effectively alleviate the mutual inference of multiple objects and improve the performance in various visual analysis tasks. To this end, we propose a novel auto-grouped sparse representation (ASR) method. ASR groups semantically correlated feature elements together through optimally fusing their multiple sparse representations. Due to the intractability of primal objective function, we also propose well-behaved convex relaxation and smooth approximation to guarantee obtaining a global optimal solution effectively. Finally, we apply ASR in two important visual analysis tasks: multi-label image classification and motion segmentation. Comprehensive experimental evaluations show that ASR is able to achieve superior performance compared with the state-of-the-arts on these two tasks.