数字X光医学影像压缩技术  

Computed X radiography image compression

作　　者：张雷[1] 杨润魁 黄廉卿[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]驻航天二院军代表室,北京100080

出　　处：《光学精密工程》2004年第6期632-637,共6页Optics and Precision Engineering

基　　金：中国科学院二期创新资助项目(No.C02E06Z)

摘　　要：针对数字X光影像仪(CR)影像的特点以及CR成像与成像板(IP)密切关系,提出一种基于图像阈值分割,并对分割的不同区域采用不同压缩方法的数字X光(CR)医学影像压缩技术。首先分割出暗背景区,进行中值滤波,然后行程编码,再根据IP板与X光曝光剂量的关系,将感兴趣区12bit像素降低量化级别降低为9.3bit来降低数据量,并采用嵌入式小波变换方法实现近无损压缩。实验结果表明,在高压缩比条件下,感兴趣区仍具有较高的峰值信噪比(PSNR),保征了感兴趣区的图像质量,具有良好的视觉效果。According to the characteristics of computed radiography (CR)imaging and the close relation ship between CR imaging procedure and image plate (IP), the compression algorithm is presented to segment image by threshold value, and different compressing methods are used for different segmented area. First, the segmentation process identifies the black surrounding and utilizes run-length coding. Then according therelation ship between imaging plate (IP) and dose of X radiography, the region of interest (ROI) CR images are transformed from 12 bit into 9.3 bit images and followed by Embedded image coding using embedded zerotree of wavelet(EZW) loss compression. This coding algorithm can ensure the image quality of the region of interest and PSNR, and give better visual effects.

关 键 词：数字化放射图像 图像压缩 感兴趣区 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]