F-TPP:TPP Image Compression with Fuzzy Resoning

Abstract

In this paper, we propose an improvement of the digital image compression using Triangular Plane Patches(TPP method). TPP method uses two triangular plane patches to approximate original image's luminance curved surfaces of a block in xy-plane. Since triangular plane patches are constructed according to the original image's luminace values of 4 vertical pixels of the block, they could hardly be suitable ones. Addressing to this problem of TPP methods, we use fuzzy reasoning to adjust the luminance values of 4 vertical pixels of a block by considering the influnce of some other pixels in the block. In this way, we can derive suitable triangular plane patches for a block by using the adjusted 4 luminace values. Our method has been implemented and experiment results show that the performance of the F-TPP method is better than that of the TPP method.