The computational cost in medical image reconstruction and processing is high, so performance improvement is required. We implemented a distributed system based on replicated worker pattern. The JINI technology was used to support a set of distributed services. The system was design and implemented to enable each replicated worker (compute server) to solve any task described by JavaSpaces Entry interface. Each worker can be configured to start as a standalone process or as a screen saver utility. The SaverBeans API was used to automatically build the screen saver code for Windows and Linux platforms. The distributed system was successfully applied to process data for medical image synthesis purposes. The achieved results proved the computational performance increase with limited modification of the original image reconstruction source code.