Lossless image compression via multi-scanning and adaptive linear prediction

This work presents an efficient lossless compression approach for gray scale images. The main contributions are: 1) we divide the input image to blocks of pixels and use different scanning methods, we select the one that minimizes the prediction error for coding; and 2) a new adaptive linear prediction is applied. Experiment results show that our method outperformed JPEG-LS 2-10% in compression efficiency.