A Study on the EMG Signal Classfier Using HMM and GA-MLP4

초록

The medical image is converting to digital format due to development of information processing and high technology. The compression is needed for image's efficient storing and transferring. In medical image, specialized for diagnosis of disease, the lossless compression scheme is preferred t o the lossy compassion scheme. Methods In this paper, We studied two image characteristics, smoothness and similarity, which give ris e to local and global redundancy in image representation. The smoothness means that the gray level val ues within a given block vary gradually rather than abruptly. The similarity means that any patterns in the image repeat itself anywhere in the rest of image. In this sense, we propose a lossless medical image compression scheme which exploits both types of red undancy. The proposed method segments the image into variable size blocks and encodes them depending on charact eristics of the block. Results The proposed compression schemes works better 10%-40% than other compression schemes such as t he Huffman, the arithmetic, the Lempel-Ziv and the lossless scheme of JPEG with one predictor. Conclusion General block coding methods concentrate on global redundancy but we focus on both global a nd local redundancy. We improved execution time and increased compression ratio.