Topology and Local Optima in Computer Vision

Abstract

We present an application of persistent homology to the image correspondence problem, also known as image registration, which is used to produce 3D reconstructions of scenery from two or more cameras. We present a novel filtered complex in the sense of persistent homology, and show that nontrivial homology groups in its persistence diagrams correspond to recognizable anomalies in images pairs, such as repeated patterns, which contribute to non-convexity of the relevant cost function. We present examples with actual image pairs, and prove a basic result that the corresponding homology classes are invariant under certain continuous deformations.

Document Details

Document Type
Pub Defense Publication
Publication Date
Jan 20, 2022
Source ID
10.1007/s42979-021-01003-x

Entities

People

  • Erik Carlsson
  • John Carlsson

Organizations

  • Defense Advanced Research Projects Agency
  • National Science Foundation
  • National Science Foundation Directorate for Mathematical & Physical Sciences

Tags

Fields of Study

  • Computer science

Readers

  • Computer Vision.
  • Mathematical Modeling and Probability Theory.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms