Processing Very Large Image Volumes with the Adaptive Particle Representation

Explore cutting-edge techniques for processing massive image datasets in this 50-minute NHR PerfLab seminar presented by Prof. Dr. Ivo Sbalzarini from the Max Planck Institute of Molecular Cell Biology and Genetics. Dive into the Adaptive Particle Representation (APR), an innovative alternative to pixel-based image representation that significantly reduces storage requirements and accelerates downstream computations. Learn how APR exploits image sparsity to adapt signal sampling, enabling image processing rates exceeding 1 TB/s on a single consumer GPU. Discover real-world applications in systems biology, including whole-brain imaging at cellular resolution, where APR has dramatically reduced processing times from days to hours. Gain insights into the challenges of visualizing, storing, and analyzing Terabyte to Petabyte-sized microscopy image datasets in developmental biology and how APR addresses these issues. Access accompanying slides and explore the speaker's background in computational science, machine learning, and high-performance computing for image-based computational biology.