Studies On Compression And Edge Detection Of Medical Image

Large amounts of digital image data, resulting in a huge consumption of existing storage media, and to the tremendous data transmission difficult. In order to achieve image storage and transmission, the key technology is image compression. For medical images, physicians are only interested in the lesion area, specified the region of interest of whole pictures, used different compression methods to different regions. Edge is one of the basic features of the image, an image is a information system, a lot of information is provided by the edge of its outline. The cell edge detection in medicine has a very large significance.In this paper, wavelet transform, morphological, pulse coupled neural networks, ant colony algorithms used in image processing, and research for specific medical image processing program specifically The main contents include the following aspects:1. Study of the method of medical image ROI compression based on lift scheme wavelet and PCNNIn order to provide a more efficient compression algorithm of medical images, lossless Regions of Interest (ROI) and lossy regions of non-interest are combined for researching. An image compression algorithm based on integer lifting wavelet transform and Pulse Coupled Neural Network (PCNN) has proposed. First, the ROI part is selected in the original image by man, second it is segmented in difference image method. Second, the ROI received is transformed by the integer 5/3 lifting wavelet and the Huffman coding.Third, the regions whose gray value similar of non-ROI are segmented by PCNN, and then use the run-length encoding. Finally, eliminate edge discontinuous of recovery image by linear interpolation.Compare with the original image, the ROI part has not been distorted and the pathological details have been preserved.By using this algorithm, a restore image with very high compression ratio and high quality in the ROI part can be obtained.2. Study several methods of an image compression algorithm based on image segmentation by improved PCNNCombine with research on human vision system, an image compression algorithm which first segment the image by improved PCNN is been proposed in this paper. The algorithm first fully adaptive segmentation using improved PCNN, the image is divided into a high level of interest regional and low information area according to the human eye. Then with the mask to distinguish, encoding them through the SPIHT respectively after the lifting wavelet transform. Through control coding bit stream, image quality and compression ratio of different regions are reasonable coordinated. Finally we using inverse transform to recover image. Compared with the original image, restore image distortion is very small in the interest regional of the human eye. Experimental results show that the realization of a relatively high compression ratio and the reconstructed image has high visual quality.3. Study a medical cell image edge detection algorithm based on morphology, PCNN and Ant colony algorithmAccording to the characteristics, the lager gradients of grey level distribution near the edges of medical cell image, the lager number of noise points, combined with morphology can be effectively de-noising and the advantages of extracting image feature points, the medical cell image edge detection algorithm based on morphology, PCNN and ant colony algorithm is proposed in this paper. By this method, first de-noising and improving edge enhance of image by morphological, and image segmentation according to the cell edge by PCNN automatic, and then ant colony algorithm is applied to edge detection. Contrast experiment shows that this method can remove noise, enhance the image edge details, and obtain a more efficient ant colony algorithm for edge detection of cell images.Figure [29] table [2] reference [99]...