Comparing classification performance of mueller matrix parameters for diffuse materials

Lisa W. Li, Meredith Kupinski, Madellyn Brown, Russell Chipman

Research output: Contribution to journalConference articlepeer-review

1 Scopus citations

Abstract

This work compares the material classification performance of Mueller matrix polarization imaging to RGB imaging. White painted wood and white fabric samples are selected to create a classification task that is challenging for RGB imaging. A Mueller Matrix Imaging Polarimeter with a 30? full field of view is used to capture the Mueller Matrix images at nominal red, green, and blue wavelengths across multiple specular scatter angles. A Bayesian ideal observer model is used to evaluate classification performance. Performance is quantified by the Area under (AUC) the Receiver Operating Characteristic (ROC) curve. An AUC = 1 is perfect detection and AUC = 0.5 is the performance of guessing. The ensemble average AUC does not exceed 0.70 for RGB irradiance data. The ensemble average AUC for all 16 individual Mueller elements is greater than 0.95. Various combinations of Mueller matrix elements are also tested. Elements related to diattenuation and polarizance are nearly perfect classifiers for large scatter angles but the AUC minimum is 0.60 at 20?. Depolarization index is the highest performing parameter out of all tested polarization parameters for scatter angles = 70? where AUC =0.98.

Original languageEnglish (US)
Article number263
JournalIS and T International Symposium on Electronic Imaging Science and Technology
Volume2020
Issue number6
DOIs
StatePublished - Jan 26 2020
Event2020 Intelligent Robotics and Industrial Applications Using Computer Vision Conference, IRIACV 2020 - Burlingame, United States
Duration: Jan 26 2020Jan 30 2020

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Science Applications
  • Human-Computer Interaction
  • Software
  • Electrical and Electronic Engineering
  • Atomic and Molecular Physics, and Optics

Fingerprint Dive into the research topics of 'Comparing classification performance of mueller matrix parameters for diffuse materials'. Together they form a unique fingerprint.

Cite this