Medical Image Compression Using Tree-Structured Vector Quantization and Fuzzy C-Means

Compression of magnetic resonance images (MRI) has proved to be more difficult than other medical imaging modalities. In an average sized hospital, many tera bytes of digital imaging data (MRI) are generated every year, almost all of which has to be kept. Compression of medical images is currently being performed by using different algorithms. In this paper, Fuzzy Clustering Method is used for the image Tree Structure Vector Quantization (TSVQ). First, MR image is used for the feature vector. Then use this feature vector to design a classification tree by Fuzzy C-Means (FCM) algorithm to split two clusters. At every nonterminal, the centroid of the feature vectors clustered in each child node is computed to be the testing vector. At every leaf, the centroid of the training image blocks corresponding to their feature vectors falling on the same terminal node is calculated to be the codevector. All codevectors in the leaves are composed of a codebook. By doing so, the algorithm can preserve the edge of image, make good image quality, and reduce the processing time while constructing Tree Structured Codebook.