A combinatorial transparent surface modeling from polarization images

Mohamad Ivan Fanany, Kiichi Kobayashi, Itsuo Kumazawa

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

This paper presents a combinatorial (decision tree induction) technique for transparent surface modeling from polarization images. This technique simultaneously uses the object's symmetry, brewster angle, and degree of polarization to select accurate reference points. The reference points contain information about surface's normals position and direction at near occluding boundary. We reconstruct rotationally symmetric objects by rotating these reference points.

Original languageEnglish
Pages (from-to)65-76
Number of pages12
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3322
DOIs
Publication statusPublished - 2004

Fingerprint

Dive into the research topics of 'A combinatorial transparent surface modeling from polarization images'. Together they form a unique fingerprint.

Cite this