Robust Object Recognition Using Color Co-occurrence Histogram and Spatial Relations of Image Patches

In this paper, a robust object recognition system is proposed, where patch-based pyramid images and spatial relationships among patches are utilized for our object model. Specially, both color histogram (CH) and color co-occurrence histogram (CCH) are applied to obtain image features for each patch. Locations of candidate regions to be tested are decided by a particle filter in our matching process. We show that the performance of object recognition can be improved by using spatial relationships among patches. To show the validities of our proposed method, there are employed input images from various environments as test images.