Lossless medical image compression using transform domain adaptive prediction for telemedicine

In some application such as medical imaging or satellite imaging an accurate reproduction of the images is one of the essential requirement. Many existing lossless image compression method are currently using spatial correlation property of neighborhood pixel to predict the data, which limits its compression efficiently due to some structural component like edges textures and contours. These types of structural elements often do exist at various location of the image many times. So, there is a high demand of development of some efficient prediction method to overrule this limitation and overcome co-linearity problem. In this paper, we have analyzed various prediction methods to minimize prediction error. The proposed prediction method is a linear predictor which perform nonlinear switching for various structural elements. Experimental results demonstrate that the propped prediction method improves compression ratio and maintain marginal PSNR along with Minimum Absolute Error.