Coronary angiography is an X-ray examination of the heart's arteries. This is an essential technique for diagnosis of heart damages. Image sequences from digital angiography contain areas of high diagnostic interest. Loss of information due to compression for regions of interest (ROI) in angiograms is not tolerable. In this paper we present a compression algorithm for compression of angiography sequences. The algorithm separates ROI from other regions. Some of the non-ROI regions are not stored in the coding process but are reconstructed with high visual quality. Block matching, prediction methods and context modelling are parts of the proposed algorithm. The implementation results show that our algorithm is more successful in compression of the angiograms as compared to standard compression routines such as JPEG-LS and JPEG-3D.
展开▼
