Abdolrahman Attar - Department of Computer Eng University of Guilan Rasht, Iran
Reza Moradi Rad - Department of Computer Eng University of Guilan Rasht, Iran
Asadollah Shahbahrami - Department of Computer Eng University of Guilan Rasht, Iran

In lossless image compression many prediction methods are proposed so far to achieve better compression performance/complexity trade off. In this paper, concentrated on some will known and widely used methods, such as MED and GAP. This paper proposes a new gradient based method that outperforms some existing methods. This claim will be proved by testing proposed method upon a wide variety of images with different characteristics. Five special sets of images including face, sport, texture, seaside, and medical constitute our dataset.

Predective Compression, Lossless Image Compression, Predictive coding, Digital Image Processing