Improving Accuracy of Pseudo Zernike Moments using Image Interpolation

Pseudo Zernike Moments (PZMs) are very popular moments among the family of orthogonal radial moments. While several methods have been proposed to enhance accuracy, accurate PZMs computation for gray level images is still an open issue. PZMs suffer from image discretization error, geometric error and numerical integration error, which result in the degradation of the reconstructed images for high order of moments. It is observed that these errors are significant for the small images. In this paper, PZMs are computed after image interpolation on the small size images. Bi-cubic interpolation is used to increase the number of sampling points of the image. Experimental results show that the proposed method provides much improved accuracy of PZMs which provide very accurate reconstructed images, numerical stability and rotation invariance