An efficient region of interest coding for medical image compression

Efficient image storage and transmission is in great demand for medical community. Compression techniques that achieve high quality for the regions important to medical diagnosis, and high overall compression ratio, may have the potential for widespread medical acceptance. JPEG 2000 provides two ROI encoding schemes: the maxshift and the implicit ROI encoding. In this paper, we present a new ROI coding scheme that not only retains advantages, but also alleviates the drawbacks of both ROI coding methods. We evaluate our method on the CT, MR, and ultrasound images, and demonstrate its coding efficiency and advantages.