Title :
Medical images compression by neural networks
Author :
Benamrane, N. ; Dah, Z. Benahmed ; Shen, J.
Author_Institution :
Dept. of Comput. Sci., USTO, Oran, Algeria
Abstract :
This paper presents a compression method for still images, based on Kohonen´s neural network. To avoid the edge degradation caused by high compression ratio, the blocks are classified into two classes : blocks with high activity (edge blocks) and blocs with low activity. The image is divided first into blocks of 16 pixels. Each block of high activity are divided again into small blocks of 4 pixels. Blocks of high and low activity are coded separately with different codebooks. We have obtained a noticeable improvement of visual quality of all the rebuild images while keeping an important compression rate. This method has been tested on medical images.
Keywords :
data compression; image coding; medical image processing; self-organising feature maps; Kohonens neural network; codebooks; high activity blocks; low activity blocks; medical image compression; visual quality improvement; Biomedical imaging; Computer science; Degradation; Electronic mail; Image coding; Image storage; Laboratories; Neural networks; Neurons; Vector quantization;
Conference_Titel :
Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the 3rd International Symposium on
Print_ISBN :
953-184-061-X
DOI :
10.1109/ISPA.2003.1296462
