engleski

Classification and blending prediction for lossless Medical image compression

Lossless compression techniques of image data are often used for the purpose of medical image archival and transmission. The sensitivity of medical data requires this kind of compression as opposed to ordinary every-day image data for which the lossy compression methods can be used. In these applications the loss of the information in the original image data is traded with the compression ratio. Among various lossless compression techniques, predictive coding techniques have proven to be very efficient in terms of both, the compression ratio and computational efficiency. In the paper we propose a new adaptive prediction scheme based on the blending of multiple static predictors on a dynamically classified causal context of neighboring pixels. The idea of predictor blends is further expanded through the determination of blending context that changes its shape on a pixel–by–pixel basis using a simple classification technique, thus allowing the modeling of more complex image structures such as nontrivially oriented edges and the periodicity and the coarseness of textures. Proposed compression algorithms estimates the image region properties around the currently unknown pixel and adjusts itself so that the presence of detected properties affects the way the compression gain is obtained. Our lossless compression algorithm can be divided in three steps: 1. prediction of the current pixel based on the causal set of surrounding pixel elements, prediction is made through the proposed classification and blending process ; 2. the determination of the current context in which the prediction error occurs, this contextual model is used for typical bias removal of the previous step and for the entropy coding of the final prediction error ; 3. the entropy coding of the final prediction error. Proposed technique is tested on a test set of typical medical images and very promising results are obtained. The test set of images are acquired using various techniques of medical imaging such as CT and MR. The various working parameters of proposed algorithm are discussed in order to adjust the algorithm to the specific properties of medical image data. We conclude with the discussion of possible avenues for the future work and improvements in the area of predictive lossless medical image compression.

lossless compression; telemedicine

nije evidentirano