Dipp is a distributed image processing pipeline for image analysis in big data contexts. It is an EXPERIMENTAL framework, mainly used to test different distribution methods. Its main use case is astronomy, where terabyte-sized images are regularly produced. To cope with these data valumes, the framework relies heavily on the MapReduce paradigm, which is implemented by thrill. The limits of the framework are tested using images from the SDSS DR18 data set.
Instructions on building the framework can be found here. If you just want to test the application or use it in a cluster environment, follow these instructions instead.