Medical imaging involves handling of huge volumes of DICOM images. Main agenda is to compress the images without compromising on the quality of the image. In this paper, a comparative analysis of different compression techniques is made for medical DICOM images. Lossless compression based on General indexing and Huffman gave maximum compression ratio 1. 6 and 1. 85. The proposed lossy compression is based on db1, db2 wavelet at single level decomposition. The proposed technique is computationally efficient since it uses a simple algorithm, at the same time achieving good PSNR, compression ratio and bits per pixel (bpp). The PSNR achieved with the proposed algorithm is always above 54. 5db across all test images. The results obtained clearly indicate that the proposed technique preserves the tumor region, thus not affecting medical diagnosis. Thus further processing like segmentation, tumor detection and classification can be applied on these compressed images.

Tumor detection lossy compression Wavelet Huffman