Combining SIFT and Invariant Color Histogram in HSV space for Deformation and viewpoint Invariant Image Retrieval

This paper presents a novel approach to retrieve images which are taken at different viewpoints, using combined feature descriptors. The content of the image is extracted with two descriptors, Scale Invariant Feature Transform (SIFT) and Deformation and view point Invariant Color Histogram (ICH) in HSV color space. SIFT has been proven to be the most reliable descriptor for rotation, translation and partially to illumination and affine or 3D projection invariant image matching. However, it is designed for gray images. Invariant Histogram is developed for creating color Histogram based on color gradients which are invariant to deformation and changes in viewpoint and is developed in RGB color space. To increase the deformation and viewpoint invariance capability and thus to improve image recognition, SIFT features are combined with ICH in HSV color space for Image Retrieval. Experimental results show that robust retrieval can be achieved even for seriously occluded images.