A pattern matching technique robust to illumination changes and occlusion

This paper presents a novel pattern matching technique that is robust to both illumination changes and the occlusion problem. The robustness to changing lighting conditions is achieved by employing gradient orientation information in place of traditional image features such as intensities and gradients. The method is further made robust to the occlusion problem by introducing a voting strategy to the matching process. The proposed method is compared with three conventional template matching techniques. Simulation results show that the proposed method performs remarkably well under varying lighting conditions and is also robust to the occlusion problem.