JPEG2000 compression for scintigraphic images of metastasis of the prostatic cancer

A remote DICOM (Digital Imaging Communication in Medicine) server needs an important storage and a high-speed network of, at least, 100 Mbps. This is a limitation for the storage and for remote access to images and image processing softwares. Objective: We set up, in this study, a secondary server that saves medical images of prostatic cancer patients compressed by JPEG 2000, the new standard of image compression. This first study has to set up an intranet and a remote server. Method: We compress to JPEG2000 the images picked up in PNG and Bitmap formats from the DICOM server using the scientific software MATLAB. Results: JPEG 2000 compression save the contrast of bone scintigraphic images. For the lossless compression the results are, respectively, 1.66 and 9.86 for the PNG format and Bitmap format. The lossy compression gives a compression ratio of, respectively, 6.48 and 40.3 for the two formats. The minimum and maximum for images compressed in PNG are 2.0 and 10.5 corresponding to a Peak Signal-to-Noise Ratio (PSNR) of, respectively, 62.9 dB and 37.8 dB. That lossy compression reduces all the images at the weight of 58 Kibi-octet (Kio). Conclusion: These compression ratios work with a good interpretation of the whole body bone scintigraphic images we used. That is proved by the visual metric as well as the PSNR. It also permits the image transmission through a low speed network. This makes easy the access on patient data in the hospital and eventually from a remote location.