A proposal of pipelined image processing in a grid environment

In the bio-medical research area, image processing for observed specimen is quite important for effective analysis. For example, 3D reconstruction processing to analyze the specimen structure is required for high performance 2D image observation equipments. It is difficult to preserve enough observation time for researchers on such premium devices. So thus, real time image analysis plays important role. Furthermore, large size data analysis is also required, because data size becomes larger as the resolution of the camera furnished for each device improves. These features require much more processing power and storage space. To realize a real time image processing for the stream data from each device, effective data staging architecture is required. In this paper, we propose architecture to handle stream data processing in a grid environment. We also propose a data staging optimization method using data partitioning.