Hydra Image Processor (also know as Hydra or HIP) is a image filtering toolkit that uses GPU acceleration.
Light microscopes can now capture data in five dimensions at very high frame rates producing terabytes of data per experiment. Five-dimensional data has three spatial dimensions (x, y, z), multiple channels ($\lambda$) and time (t). Current tools are prohibitively time consuming and do not efficiently utilize available hardware. The Hydra is a new library providing hardware-accelerated image processing accessible from interpreted languages including MATLAB and Python. Hydra automatically distributes data/computation across system and video RAM allowing hardware-accelerated processing of arbitrarily large images. Hydra also partitions compute tasks optimally across multiple GPUs. Hydra includes a new kernel renormalization reducing boundary effects associated with widely used padding approaches.