Gpu-based image processing for multiview microscopy data

Multi-view light-sheet microscopy is one of the best tools to investigate early embryonic development due to its inherent optical sectioning, fast imaging speed, and low phototoxicity. Evaluating, or simply just storing the data, however, is an extremely challenging step, since each experiment can generate terabytes of data. Furthermore, this data has to be pre-processed before any evaluation step can be performed, i.e. multiple views have to be fused to create a single, high quality image.