The Slow Road to Fast Persistent Homology for Image Data

Explore fifteen years of computational advances in persistent homology for image data analysis through this 59-minute conference talk by Hubert Wagner. Discover the evolution of efficient algorithms and techniques developed to overcome computational obstacles in applying persistent homology to image datasets. Learn about the key breakthroughs that have enabled faster computations, examine the current software ecosystem available for these calculations, and understand the realistic data size limitations when working with modern hardware including GPU implementations. Gain insights into the practical considerations for deploying persistent homology methods in real-world image analysis applications and understand how the field has progressed from slow, computationally intensive methods to more efficient approaches suitable for contemporary data processing needs.