A Novel Image Retrieval Approach for Digital Images Based On Pseudo-Zernike Moment Invariants

In this paper we propose a new and effective image retrieval scheme using color, texture and shape features based on pseudo-Zernike moment. Image is predefined by using fast color quantization algorithm with cluster merging. A small number of dominant colors and their percentage can be obtained. The spatial features are extracted using steerable filter decomposition. This offers an efficient and flexible approximation of early processing in the human visual system (HVS). Sharp descriptor of an image uses pseudo-Zernike moment. It provides better feature representation and more robust to noise then other representations. Finally, the combination of the color, shape and texture feature provides a robust feature set for image retrieval. Experimental result shows that the proposed method provides a better color image retrieval. It provides accurate and effective in retrieving the user-interested images.