attached
Medical image processing has become a major sub area in digital image processing which has a wide
range of applications in which image compression has become a core sub area of medical image
processing due to the wide and frequent usage of medical images all over the world. The high quality
of the images are very important for diagnosis and on the other hand, they require huge memory,
bandwidth and storage space for processing, transmission and archival in digital form.
This research presents an approach which compresses 2D human brain images which are in DICOM
format in a hybrid manner using both lossy and lossless methods. This approach basically depends
on the wavelet transformation where The regions which are clinically very important can be
compressed in lossy manner while others are compressed in a visually lossless manner using the
Huffman coding algorithm.
Results indicate that compression of large images tend to show better compression ratio than the
small sized images. The best compression ratio achieved using the proposed algorithm is 0.84 for
grey scaled 2D DICOM images. Resultant compressed image and header data can be stored or
transmitted through a computer network effectively to another destination for retrieval.
