CT图像混合压缩算法研究

Study on hybrid compression algorithm of CT images

计算机断层扫描图像数据量往往很大,不利于存储和传输。一种CT图像混合压缩算法被提出,利用CT序列图像间存在相关性,以第一幅图像为参考图像,将后续图像和参考图像进行差值运算,消除图像序列间冗余。同时对图像的感兴趣区域采用哈夫曼无损压缩,对非感兴趣区域采用类似于JPEG基线模式的有损压缩。实验结果表明,新提出的算法在保证CT图像质量的前提下,能够有效地提高压缩比,具有一定的临床应用价值。

It is necessary that compress computed tomography （CT） images which size are usually huge against storage and transmission. A hybrid compression algorithm of CT images was proposed. The strong correlation between CT set images was utilized, that is the first one in CT set images was regarded as reference image and the differences between successive ones and reference one were made in order to reduce redundancy. At the same time, region of interest （ROI） in image was compressed using Huffman lossless coding and other regions were compressed using a lossy coding similar to JPEG baseline mode. The experimental results indicate that our algorithm could improve the compression ratio without significant losses in the image quality and benefit clinical practice application.