A Multi-Parameter Persistence Framework for Mathematical Morphology
Offered By: Applied Algebraic Topology Network via YouTube
Course Description
Overview
Explore a 55-minute lecture on applying persistent homology to mathematical morphology for image processing. Delve into the integration of topological data analysis with classic morphological operations, forming multiparameter filtrations. Learn how this framework extracts topological and geometric information from images, enabling automated optimization of image structure analysis and rendering. Discover an unsupervised denoising approach for binary, grayscale, and color images, comparable to state-of-the-art supervised deep learning methods. Cover topics including opening and closing operations, shifting functions, persistence diagrams, multiparameter filtrations, sublevel sets, and computational complexity. Gain insights into the potential of this innovative framework for enhancing image processing techniques through the lens of applied algebraic topology.
Syllabus
Introduction
Background knowledge notation
What is mathematical morphology
Opening and closing operations
New filtration
Shifting function
Persistence diagram
Variants
Multiparameter filtration
Alternating closing
Alternating opening
Assumption
Sublevel set
Basic idea
Comparison
Summary
Questions
Algebraic structure
Randomization
Natural biofiltration
Opening and closing
Computational complexity
Short answer
Taught by
Applied Algebraic Topology Network
Related Courses
Topology for Time SeriesData Science Dojo via YouTube Studying Fluid Flows with Persistent Homology - Rachel Levanger
Institute for Advanced Study via YouTube Persistence Diagram Bundles- A Multidimensional Generalization of Vineyards
Applied Algebraic Topology Network via YouTube GPU Accelerated Computation of VR Barcodes in Evaluating Deep Learning Models
Applied Algebraic Topology Network via YouTube New Results in Computing Zigzag and Multiparameter Persistence
Applied Algebraic Topology Network via YouTube