Information-theory-based registration and segmentation of medical images
Karol Kuczyński, Paweł Mikołajczak · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2004
Growing popularity of non invasive medical techniques makes accurate and reliable diagnostics based on image data more and more important. It is necessary to implement modern, highly automated image processing techniques. Information theory and statistics provide a means to create such systems. In this paper we discuss application of entropy and mutual information in segmentation and registration of real world medical images.