Nonparametric Bayesian Dictionary Learning for Analysis of Noisy and Incomplete Images

Abstract

Nonparametric Bayesian methods are considered for recovery of imagery based upon compressive incomplete and/or noisy measurements. A truncated beta-Bernoulli process is employed to infer an appropriate dictionary for the data under test, and also for image recovery. In the context of compressive sensing, significant improvements in image recovery are manifested using learned dictionaries, relative to using standard orthonormal image expansions. The compressive-measurement projections are also optimized for the learned dictionary. Additionally, we consider simpler (incomplete) measurements defined by measuring a subset of image pixels, selected uniformly at random; connections are made to matrix completion and union-of-subspace models, providing a link between matrix completion and image processing. Spatial inter-relationships within imagery are exploited through use of the Dirichlet and probit stick-breaking processes. Several example results are presented, with comparisons to other methods in the literature.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 2010
Accession Number
ADA571911

Entities

People

  • David B. Dunson
  • Guillermo Sapiro
  • Haojun Chen
  • John Paisley
  • Lingbo Li
  • Lu Ren
  • Mingyuan Zhou
  • Zhengming Xing

Organizations

  • University of Minnesota

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Bayesian Networks
  • Compressed Sensing
  • Databases
  • Dictionaries
  • Factor Analysis
  • Gray Scale
  • Image Processing
  • Image Segmentation
  • Information Processing
  • Learning
  • Measurement
  • Monte Carlo Method
  • Numbers
  • Probability
  • Recovery
  • Standards

Readers

  • Image Processing and Computer Vision.
  • Regression Analysis.
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference