Belief propagation with directional statistics for solving the shape-from-shading problem

Tom Fincham Haines, Richard C Wilson

Research output: Contribution to conferencePaper

6 Citations (Scopus)

Abstract

The Shape-from-Shading [SfS] problem infers shape from reflected light, collected using a camera at a single point in space only. Reflected light alone does not provide sufficient constraint and extra information is required; typically a smoothness assumption is made. A surface with Lambertian reflectance lit by a single infinitely distant light source is also typical.
We solve this typical SfS problem using belief propagation to marginalise a probabilistic model. The key novel step is in using a directional probability distribution, the Fisher-Bingham distribution. This produces a fast and relatively simple algorithm that does an effective job of both extracting details and being robust to noise. Quantitative comparisons with past algorithms are provided using both synthetic and real data.

Conference

ConferenceEuropean Conference on Computer Vision
Abbreviated titleECCV
CountryFrance
CityMarseille
Period12/10/0818/10/08

Fingerprint

Statistics
Probability distributions
Light sources
Cameras
Statistical Models

Cite this

Fincham Haines, T., & Wilson, R. C. (2008). Belief propagation with directional statistics for solving the shape-from-shading problem. 780-791. Paper presented at European Conference on Computer Vision, Marseille, France.

Belief propagation with directional statistics for solving the shape-from-shading problem. / Fincham Haines, Tom; Wilson, Richard C.

2008. 780-791 Paper presented at European Conference on Computer Vision, Marseille, France.

Research output: Contribution to conferencePaper

Fincham Haines, T & Wilson, RC 2008, 'Belief propagation with directional statistics for solving the shape-from-shading problem' Paper presented at European Conference on Computer Vision, Marseille, France, 12/10/08 - 18/10/08, pp. 780-791.
Fincham Haines T, Wilson RC. Belief propagation with directional statistics for solving the shape-from-shading problem. 2008. Paper presented at European Conference on Computer Vision, Marseille, France.
Fincham Haines, Tom ; Wilson, Richard C. / Belief propagation with directional statistics for solving the shape-from-shading problem. Paper presented at European Conference on Computer Vision, Marseille, France.12 p.
@conference{f9967c72fd964d06bf951a403a846923,
title = "Belief propagation with directional statistics for solving the shape-from-shading problem",
abstract = "The Shape-from-Shading [SfS] problem infers shape from reflected light, collected using a camera at a single point in space only. Reflected light alone does not provide sufficient constraint and extra information is required; typically a smoothness assumption is made. A surface with Lambertian reflectance lit by a single infinitely distant light source is also typical.We solve this typical SfS problem using belief propagation to marginalise a probabilistic model. The key novel step is in using a directional probability distribution, the Fisher-Bingham distribution. This produces a fast and relatively simple algorithm that does an effective job of both extracting details and being robust to noise. Quantitative comparisons with past algorithms are provided using both synthetic and real data.",
author = "{Fincham Haines}, Tom and Wilson, {Richard C}",
year = "2008",
language = "English",
pages = "780--791",
note = "European Conference on Computer Vision, ECCV ; Conference date: 12-10-2008 Through 18-10-2008",

}

TY - CONF

T1 - Belief propagation with directional statistics for solving the shape-from-shading problem

AU - Fincham Haines, Tom

AU - Wilson, Richard C

PY - 2008

Y1 - 2008

N2 - The Shape-from-Shading [SfS] problem infers shape from reflected light, collected using a camera at a single point in space only. Reflected light alone does not provide sufficient constraint and extra information is required; typically a smoothness assumption is made. A surface with Lambertian reflectance lit by a single infinitely distant light source is also typical.We solve this typical SfS problem using belief propagation to marginalise a probabilistic model. The key novel step is in using a directional probability distribution, the Fisher-Bingham distribution. This produces a fast and relatively simple algorithm that does an effective job of both extracting details and being robust to noise. Quantitative comparisons with past algorithms are provided using both synthetic and real data.

AB - The Shape-from-Shading [SfS] problem infers shape from reflected light, collected using a camera at a single point in space only. Reflected light alone does not provide sufficient constraint and extra information is required; typically a smoothness assumption is made. A surface with Lambertian reflectance lit by a single infinitely distant light source is also typical.We solve this typical SfS problem using belief propagation to marginalise a probabilistic model. The key novel step is in using a directional probability distribution, the Fisher-Bingham distribution. This produces a fast and relatively simple algorithm that does an effective job of both extracting details and being robust to noise. Quantitative comparisons with past algorithms are provided using both synthetic and real data.

M3 - Paper

SP - 780

EP - 791

ER -