In this paper

we propose an efficient algorithm for compressing gray-level finger-print image.This algorithm is implemented by combining signal-noise characteristic of finger-print image with context-based prediction coding

the baseline algorithm for next generation of international standard on lossless image compression.There are mainly three distinguishing features in our proposed algorithm:local direction-based prediction

extended context for micro texture and histogram initialization based on imaging apparatus.Experiments with real finger print images and ISO standard images show that the compression ratio of our method is among the best in the literature.