Role of visual languages in developing image analysis algorithms

Developers of new image analysis algorithms typically require an interactive environment in which the image data can be passed through various operators, some of which may involve feedback, synchronization, merging and conditional control strategies. This paper describes the requirements of visual languages for developing new algorithms, and how several visual languages meet these requirements. We then show how the data flow methodology of a visual software development tool called WIT has greatly simplified the prototyping and testing of our new image registration and segmentation methods. We describe the main features of WIT, and show data flow graphs for two medical image analysis algorithms