Medical Images Adaptive Compression With Statistical Model for Transmission and Archiving; Application to MRI modality

Image compression with loss presents a major interest or profit for archiving and image transmission. Many methods are used in this field. The most useful is based on the Discret Cosine Transform DCT. It then, represents the core of the JPEG's norm on the compression of fixed images. Medical structures involue an important volume of administrative and medical information where imagery could be very high highted. But the quantity of information represented by images causes serious problems. Our task deals with algorithmic aspects looking for compression technique adapted to different modalities of medical imagery and on the validity of images. A statistical study has been extensively done so as to determine the laws which follow the coefficients resulting from DCT. The results which are already obtained will be used to work with during or when compressing. The validity of diagnostic quality of medical images is largely based on an assessment made by a group of medical experts. We finally come to observe that the diagnostic quality of the image is kept. Globally, the results are very encouraging and positive they eventually show that compression is acceptable and largely useful in medical work. Key-words: Compression; Discret Cosine Transform DCT; JPEG; MRI; Adapted quantification; Diagnostic quality.