A REGION BASED PREDICTOR FOR LOSSLESS COMPRESSION OF RNAI IMAGES

In this paper a new prediction method is proposed for compression of RNAi images. The large number of RNAi images that are produced, from experiments on biological cells for diagnosis and prognosis of diseases, require special compression methods. Images are segmented so that the boundaries of the cells are recognized from the smooth areas. The proposed scheme adaptively changes its function to exploit the spatial features of the cell boundaries and the smooth regions. The proposed predictor has either better performance and comparable complexity, or it has lower complexity and comparable performance, when compared to the existing predictors for this specific application.