Research Plan Proposals

Development of the Simultaneous Processing System to
Combine, Imprint and Compress Medical Images Distributed Through the Internet

Masahiro NISHIBORI^*1 (mn-mn@umin.ac.jp)
Hiroshi TANAKA^*2, Borut ZALIK^*3
*1Clinical Laboratory, Tokyo Medical and Dental University Medical Hospital
^*2Information Center, Tokyo Medical and Dental University
^*3Faculy of Electrical Engineering and Computer Science, University of Maribor

1. Objectives of the project
The distribution of digitized medical images through the network is essential for the telemedicine, electronic patient record management and remote medical education. While the development of special networks for these tasks only does not seem realistic, it is becoming more and more natural to use capabilities offered by internet. Unfortunately, some serious limitations are encountered in using the internet for medical use, and they must be overcome first. The attention should be paid to improve the transmission rate, to increase security of transmitted data and to standardize image formats on both, sending and receiving side, while the images should retain their diagnostic value after the transmission.
The outcome of the project will be the working computer program with all necessary functionalities. It is anticipated that the program would probably not be optimized regarding the speed and the rate of the compression in this phase of cooperation. The results of the project will be published as common papers in international conferences and international journals. Specifically, compressing the textual and the imaging information mixed in the common file seems to be new and unique solution.

2. Background for the collaboration
These tasks require cooperation of both computer science and medical staff. While the best cognition of standard techniques of image and text compression and encryption and the ability of combining and upgrading them is being expected from the computer science experts and programmers, the medical team is more than anyone else competent to evaluate the consistency and the speed of the transmission process. After all, only a medical group can provide the initial data and the knowledge of the medical images and medical text information, and as final users they can also do all necessary testings in the most efficient way. In our case, The Japanese side consists of the group led by Dr. Masahiro NISHIBORI, coming from a medical field and specialized in using medical images through the internet. The Slovene group led by Dr. Borut ZALIK comes from the field of computer science and is specialized in computer graphics, computational geometry and multimedia algorithms.

3. Scientific and technical description of the project including methodology
The internet has some limitations for medical use:
(1) the equipment of input and output for image data is not standardized.
(2) neither the owner nor the sender of data cannot be traced.
(3) consistency between sent data and received data is not guaranteed.
(4) the capacity of data transmission is various and limited.
To overcome these limitations, the following techniques can be used:
(1) the specifications of input equipment and the parameters of input process are attached to images and the output process is adjusted using these information.
(2) a secure server and an electronic identification are used for authorization.
(3) some special data format such as DICOM has been provided for secure transmission of image data.
(4) various methods for data compression are used to save the data load on the network.
At present, these devices are equipped separately, and are not always compatible with each other. We need a new system in which every function is more seamlessly provided, no special equipment is required and the diagnostic value of each image after compression is guaranteed. This system will combine an image file with attached information such as copyright, conditions for data output and diagnostic comments put into a file, imprint an electronic signature which affirms the true sender so that any unauthorized modification is detected, and compress images to the minimum size so as not to lose their diagnostic values, and will also give the reverse functions of them.
To realize these goals, some standard techniques known from the theory of still imaging compression, text compression and encryption will be combined and upgraded. Standard JPEG (Joint Photographic Expert Group) for example ensures the compression rates up to 10 without noticeable loss in the quality of the restored image. Known techniques and algorithms developed for the standard JPEG up to the final level - the coding - will be used. A special coding scheme will be developed ensuring that only the persons with decoding software developed by us would be able to see the image. For the text compression, the adaptive Huffman coding algorithm will be implemented. The text codes and the picture information will be mixed in the same file making the decompression by an unauthorized person even more difficult. The most crucial textual information (such as electronic identification) will be protected by the RSA encryption algorithm. Only with the correct identification (the principle of the public keys), decompression will be possible.

4. Possible collaborators to be invited
An application for academic grants related to this project has been made to Japanese government. Researchers interested in this project as well as manufacturers willing to produce the system and able to offer essential resources required to develop it are welcome.