Supporting Zooming-in Process for Image Compression Based on High-Order Weighted 3D Polynomials Fitting


Abstract –This paper presents a proposed technique to compress images using weighted 3D polynomials fitting technique that fits all pixels as possible in the image. This technique uses high-order weighted 3D polynomials to obtain high quality compressed images especially in the medical images. These types of images seek for high details with an acceptable compression ratio. This procedure of weighted 3D polynomials fitting ensures to preserve the quality of image during the decompression and zooming-in process. After applying scalar quantization and Huffman encoding to the weighted polynomials coefficients for each block of image; mean square error (MSE), peak signal to noise ratio (PSNR), processing time, and compression ratio (CR) are evaluated for different degree of weighted polynomials and for different medical image block sizes. Computer results showed that the proposed technique gives an acceptable image quality under zooming-in process compared with standard surface fitting that uses non-weighted polynomials but at the expanse of compression ratio.