报告摘要:
Image feature matching is a fundamental problem in computer vision area. Feature matching problem is usually formulated as an Integer Quadratic Programming (IQP) problem. Since it is NP-hard, relaxation models are required to find approximate solutions. In this talk, we will present our recent works on developing some sparse relaxation models for feature matching problem. The aim of the proposed matching models is to find an optimal sparse solution (approximate discrete solution) for feature matching problem. We derive a simple yet effective update algorithms to solve the proposed models.