Common Visual Patterns Discovery with an Elastic Matching Model

Meili Zhao, Bo Jiang, Bin Luo, Jin Tang

Abstract: Common visual patterns discovery (CVP) is a fundamental problem in the computer vision area. It has been widely used in many computer vision tasks. Recent works have formulated this problem as a dense subgraph detection problem. Since it is NP-hard, approximate methods are required. In this paper, we propose a new method for CVP problem, called Elastic Matching (ElasticM). The main feature of the proposed ElasticM is that it uses $$\ell _p$$ ℓ p norm constraint to induce sparse solution and thus conducts detection task naturally and more robustly in its optimization process. Promising experimental results demonstrate the benefit of the proposed CVP discovery method.