An Effective Optimization of Reference Pixel Arrangement in the Lossless Coding for HDTV image

This paper presents a lossless image compression method for one frame of High-definition television (HDTV). We apply classified adaptive prediction, and then the prediction error is encoded by entropy coding of arithmetic coding. Then an image is divided into small blocks, and they are classified into some classes each of which correspond to one minimum mean square error (MMSE) linear predictor. In this paper, we consider the influence of the reference pixel of context and the number of the linear predictors for the compression ratio and encoding time. After the fast optimization of block classification with a small number of reference pixels of context, we optimize the number, and the arrangement of reference pixels based on variable selection technique which uses F-statistic (test statistic of partial regression coefficient) in multiple linear regression analysis. In this paper, we propose an effective optimization method of reference pixel arrangement in the lossless coding for HDTV still image. As a result of experiments, it is confirmed that our method achieves higher compression ratios and faster processing speed, than conventional one. It should be noted that the decoding speed of the proposed method is very fast.