Deakin University
Browse

File(s) under permanent embargo

Spectral dichromatic parameter recovery from two views via total variation hyper-priors

conference contribution
posted on 2017-01-01, 00:00 authored by Filippo Bergamasco, Andrea Torsello, Antonio Robles-KellyAntonio Robles-Kelly
In this paper, we propose an approach for the recovery of the dichromatic model from two hyperspectral or multispectral images, i.e., the joint estimation of illuminant, reflectance, and shading of each pixel, as well as the optical flow between the two views. The approach is based on the minimization of an energy functional linking the dichromatic model to the image appearances and the flow between the images to the factorized reflectance component. In order to minimize the resulting under-constrained problem, we apply vectorial total variation regularizers both to the scene reflectance, and to the flow hyper-parameters. We do this by enforcing the physical priors for the reflectance of the materials in the scene and assuming the flow varies smoothly within rigid objects in the image. We show the effectiveness of the approach compared with single view model recovery both in terms of model constancy and of closeness to the ground truth.

History

Volume

10116

Pagination

317-333

Location

Taipei, Taiwan

Start date

2016-11-20

End date

2016-11-24

ISBN-13

978-3-319-54407-6

Language

eng

Publication classification

E1.1 Full written paper - refereed

Copyright notice

2017, Springer International Publishing AG

Editor/Contributor(s)

Chen CS, Lu J, Ma KK

Title of proceedings

ACCV 2016 : Proceedings of the 13th Asian Conference on Computer Vision

Event

Asian Federation of Computer Vision. Conference (13th : 2016 : Taipei, Taiwan)

Publisher

Springer

Place of publication

Cham, Switzerland

Series

Asian Federation of Computer Vision Conference

Usage metrics

    Research Publications

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC