On image segmentation using information theoretic criteria

DOI10.1214/11-AOS925zbMATH Open1246.94012arXiv1203.2087OpenAlexW1980384004MaRDI QIDQ449974FDOQ449974

Authors: Alexander Aue, Thomas C. M. Lee 

Publication date: 3 September 2012

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Image segmentation is a long-studied and important problem in image processing. Different solutions have been proposed, many of which follow the information theoretic paradigm. While these information theoretic segmentation methods often produce excellent empirical results, their theoretical properties are still largely unknown. The main goal of this paper is to conduct a rigorous theoretical study into the statistical consistency properties of such methods. To be more specific, this paper investigates if these methods can accurately recover the true number of segments together with their true boundaries in the image as the number of pixels tends to infinity. Our theoretical results show that both the Bayesian information criterion (BIC) and the minimum description length (MDL) principle can be applied to derive statistically consistent segmentation methods, while the same is not true for the Akaike information criterion (AIC). Numerical experiments were conducted to illustrate and support our theoretical findings.

Full work available at URL: https://arxiv.org/abs/1203.2087

Recommendations

• Robust image segmentation via Bayesian type criterion
• scientific article; zbMATH DE number 6311561
• A Minimum Description Length-Based Image Segmentation Procedure, and Its Comparison with a Cross-Validation-Based Segmentation Procedure
• scientific article; zbMATH DE number 431941
• scientific article; zbMATH DE number 1070495

zbMATH Keywords

statistical consistencyAkaike information criterion (AIC)Bayesian information criterion (BIC)image modelingminimum description length (MDL)piecewise constant function modeling

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Applications of statistics in engineering and industry; control charts (62P30) Filtering in stochastic control theory (93E11) Image analysis in multivariate analysis (62H35) Image processing (compression, reconstruction, etc.) in information and communication theory (94A08)

Cites Work

• Estimating the dimension of a model
• A new look at the statistical model identification
• Estimating the number of change-points via Schwarz' criterion
• A limit theorem for the maximum of normalized sums of independent random variables
• Information and complexity in statistical modeling.
• Title not available (Why is that?)
• Estimating the Number of Change Points in Exponential Families Distributions
• An Edge-Weighted Centroidal Voronoi Tessellation Model for Image Segmentation
• Title not available (Why is that?)
• A Minimum Description Length-Based Image Segmentation Procedure, and Its Comparison with a Cross-Validation-Based Segmentation Procedure
• Title not available (Why is that?)

Cited In (11)

• Testing for common breaks in a multiple equations system
• Structural breaks in time series
• Detecting abrupt changes in the spectra of high-energy astrophysical sources
• Robust image segmentation via Bayesian type criterion
• Title not available (Why is that?)
• Automatic estimation of flux distributions of astrophysical source populations
• Rejoinder on: ``Extensions of some classical methods in change point analysis
• Alternating Pruned Dynamic Programming for Multiple Epidemic Change-Point Estimation
• Local bandwidth selection via second derivative segmentation
• Segmented model selection in quantile regression using the minimum description length principle
• Title not available (Why is that?)

This page was built for publication: On image segmentation using information theoretic criteria

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q449974)