Journal of Data Processing Suraya Abu Bakar, Muhammad Suzuri Hitam, Wan Nural Jawahir Hj Wan Yussof 3 1 2013 http://www.dline.info/jdp/fulltext/v3n1/2.pdf An important issue to be solved in object recognition is how to represent an object so that a computer vision system is able to recognize it regardless of their shapes, sizes, position and orientation. In this paper we propose a shape based image retrieval based on Zernike Moments as shape descriptor for retrieval of single binary object from image database. To test the robustness of Zernike Moments, MPEG-7 image database is used. Our experimental investigations confirm that Zernike Moment properties are invariant to rotation, and translation with low sensitivity to noise, however is not invariant to scale.