一种基于医学序列图像的“有损-无损”压缩方案

A Novel Lossy-Lossless Compression Approach Based on Medical Image Sequence

针对医学序列图像的特点 ,提出了基于图像配准技术的差值预测方法 ,以减少序列图像之间的相关冗余。对于预测差值图像 ,采用基于可逆整数小波变换的EBCOT编码方法 ,实现对“非必要 -必要”区域的“有损 -无损”压缩。通过对实际医学图像 (CT)的实验表明 ,该方法具有很好的压缩性能 ,可以有效满足现代医学图像的压缩要求。

According to the peculiarity of Medical Image Sequence, this paper proposes a novel approach of compression. The idea is to use differential prediction based on the image registration technique to decrease the redundancies between the benchmark image and the other images in the sequence. Further more, this method uses an algorithm called EBCOT based on reversible integer wavelet transform to compress the residual differential image in the Lossy-Lossless mode. Results of experiments with real medical image (CT) help illustra-te this schema is an effective compressive encoding approach.