Design and implementation of a client-server medical image database system with image tiling and image compression

Document Type



Master of Science (MS)


Computer Science

First Advisor's Name

Nagarajan Prabakar

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Shu-Ching Chen

Third Advisor's Name

Masoud Milani

Date of Defense



The increasing amount of medical image data makes it difficult to access medical database over the Internet which is subject to a limited network bandwidth. This thesis addresses the problem of image tiling and compression, and introduces the design and implementation of a medical image database system, which provides an easy mean to access medical image data over network.

First, a Java image pre-processing module is designed and implemented to do the image tiling and compression. Image tiling allows clients to get desired image tiles over a network instead of the whole images and consequently improves response time and reduces memory requirement. Image compression, including lossy image compression method (JPG) and lossless image compression (Zip), further compresses the tiled image for saving storage space. To facilitate the image processing, medical image boundary detection and boundary alignment for the entire format set have been designed and implemented. Then, a database has been designed to maintain the medical images (tiles). A server module is also implemented to process the client requests and send image tiles back to clients. Java RMI (remote method innovation) is used for communications between servers and clients.

The system is implemented in Java language under the Visual Cafe 3.0 environment. Microsoft SQL server 7.0 is used for relational database. JDBC and JDBC-ODBC bridge are used to access the database.



This document is currently not available here.