Combining probabilistic shape-from-shading and statistical facial shape models

Touqeer Ahmad, Richard C Wilson, William A. P. Smith, Tom Fincham Haines

Research output: Contribution to conferencePaper

Abstract

Shape-from-shading is an interesting approach to the problem of finding the shape of a face because it only requires one image and no subject participation. However, SfS is not accurate enough to produce good shape models. Previously, SfS has been combined with shape models to produce realistic face reconstructions. In this work, we aim to improve the quality of such models by exploiting a probabilistic SfS model based on Fisher-Bingham 8-parameter distributions (FB8). The benefits are two-fold; firstly we can correctly weight the contributions of the data and model where the surface normals are uncertain, and secondly we can locate areas of shadow and facial hair using inconsistencies between the data and model. We sample the FB8 distributions using a Gibbs sampling algorithm. These are then modelled as Gaussian distributions on the surface tangent plane defined by the model. The shape model provides a second Gaussian distribution describing the likely configurations of the model; these distributions are combined on the tangent plane of the directional sphere to give the most probable surface normal directions for all pixels. The Fisher criterion is used to locate inconsistencies between the two distributions and smoothing is used to deal with outliers originating in the shadowed and specular regions. A surface height model is then used to recover surface heights from surface normals. The combined approach shows improved results over the case when only surface normals from shape-from-shading are used.
Original languageEnglish
Pages680-690
Number of pages11
DOIs
Publication statusPublished - 2011
EventInternational Conference on Image Analysis and Processing - Ravenna, Italy
Duration: 14 Sep 201116 Sep 2011
Conference number: 16

Conference

ConferenceInternational Conference on Image Analysis and Processing
Abbreviated titleICIAP
CountryItaly
CityRavenna
Period14/09/1116/09/11

Fingerprint

Gaussian distribution
Pixels
Sampling
Statistical Models

Cite this

Ahmad, T., Wilson, R. C., Smith, W. A. P., & Fincham Haines, T. (2011). Combining probabilistic shape-from-shading and statistical facial shape models. 680-690. Paper presented at International Conference on Image Analysis and Processing, Ravenna, Italy. https://doi.org/10.1007/978-3-642-24085-0_69

Combining probabilistic shape-from-shading and statistical facial shape models. / Ahmad, Touqeer; Wilson, Richard C; Smith, William A. P.; Fincham Haines, Tom.

2011. 680-690 Paper presented at International Conference on Image Analysis and Processing, Ravenna, Italy.

Research output: Contribution to conferencePaper

Ahmad, T, Wilson, RC, Smith, WAP & Fincham Haines, T 2011, 'Combining probabilistic shape-from-shading and statistical facial shape models' Paper presented at International Conference on Image Analysis and Processing, Ravenna, Italy, 14/09/11 - 16/09/11, pp. 680-690. https://doi.org/10.1007/978-3-642-24085-0_69
Ahmad T, Wilson RC, Smith WAP, Fincham Haines T. Combining probabilistic shape-from-shading and statistical facial shape models. 2011. Paper presented at International Conference on Image Analysis and Processing, Ravenna, Italy. https://doi.org/10.1007/978-3-642-24085-0_69
Ahmad, Touqeer ; Wilson, Richard C ; Smith, William A. P. ; Fincham Haines, Tom. / Combining probabilistic shape-from-shading and statistical facial shape models. Paper presented at International Conference on Image Analysis and Processing, Ravenna, Italy.11 p.
@conference{b6f4445e6cf44e4eb6df8e77864e8185,
title = "Combining probabilistic shape-from-shading and statistical facial shape models",
abstract = "Shape-from-shading is an interesting approach to the problem of finding the shape of a face because it only requires one image and no subject participation. However, SfS is not accurate enough to produce good shape models. Previously, SfS has been combined with shape models to produce realistic face reconstructions. In this work, we aim to improve the quality of such models by exploiting a probabilistic SfS model based on Fisher-Bingham 8-parameter distributions (FB8). The benefits are two-fold; firstly we can correctly weight the contributions of the data and model where the surface normals are uncertain, and secondly we can locate areas of shadow and facial hair using inconsistencies between the data and model. We sample the FB8 distributions using a Gibbs sampling algorithm. These are then modelled as Gaussian distributions on the surface tangent plane defined by the model. The shape model provides a second Gaussian distribution describing the likely configurations of the model; these distributions are combined on the tangent plane of the directional sphere to give the most probable surface normal directions for all pixels. The Fisher criterion is used to locate inconsistencies between the two distributions and smoothing is used to deal with outliers originating in the shadowed and specular regions. A surface height model is then used to recover surface heights from surface normals. The combined approach shows improved results over the case when only surface normals from shape-from-shading are used.",
author = "Touqeer Ahmad and Wilson, {Richard C} and Smith, {William A. P.} and {Fincham Haines}, Tom",
year = "2011",
doi = "10.1007/978-3-642-24085-0_69",
language = "English",
pages = "680--690",
note = "International Conference on Image Analysis and Processing, ICIAP ; Conference date: 14-09-2011 Through 16-09-2011",

}

TY - CONF

T1 - Combining probabilistic shape-from-shading and statistical facial shape models

AU - Ahmad, Touqeer

AU - Wilson, Richard C

AU - Smith, William A. P.

AU - Fincham Haines, Tom

PY - 2011

Y1 - 2011

N2 - Shape-from-shading is an interesting approach to the problem of finding the shape of a face because it only requires one image and no subject participation. However, SfS is not accurate enough to produce good shape models. Previously, SfS has been combined with shape models to produce realistic face reconstructions. In this work, we aim to improve the quality of such models by exploiting a probabilistic SfS model based on Fisher-Bingham 8-parameter distributions (FB8). The benefits are two-fold; firstly we can correctly weight the contributions of the data and model where the surface normals are uncertain, and secondly we can locate areas of shadow and facial hair using inconsistencies between the data and model. We sample the FB8 distributions using a Gibbs sampling algorithm. These are then modelled as Gaussian distributions on the surface tangent plane defined by the model. The shape model provides a second Gaussian distribution describing the likely configurations of the model; these distributions are combined on the tangent plane of the directional sphere to give the most probable surface normal directions for all pixels. The Fisher criterion is used to locate inconsistencies between the two distributions and smoothing is used to deal with outliers originating in the shadowed and specular regions. A surface height model is then used to recover surface heights from surface normals. The combined approach shows improved results over the case when only surface normals from shape-from-shading are used.

AB - Shape-from-shading is an interesting approach to the problem of finding the shape of a face because it only requires one image and no subject participation. However, SfS is not accurate enough to produce good shape models. Previously, SfS has been combined with shape models to produce realistic face reconstructions. In this work, we aim to improve the quality of such models by exploiting a probabilistic SfS model based on Fisher-Bingham 8-parameter distributions (FB8). The benefits are two-fold; firstly we can correctly weight the contributions of the data and model where the surface normals are uncertain, and secondly we can locate areas of shadow and facial hair using inconsistencies between the data and model. We sample the FB8 distributions using a Gibbs sampling algorithm. These are then modelled as Gaussian distributions on the surface tangent plane defined by the model. The shape model provides a second Gaussian distribution describing the likely configurations of the model; these distributions are combined on the tangent plane of the directional sphere to give the most probable surface normal directions for all pixels. The Fisher criterion is used to locate inconsistencies between the two distributions and smoothing is used to deal with outliers originating in the shadowed and specular regions. A surface height model is then used to recover surface heights from surface normals. The combined approach shows improved results over the case when only surface normals from shape-from-shading are used.

UR - https://doi.org/10.1007/978-3-642-24085-0_69

U2 - 10.1007/978-3-642-24085-0_69

DO - 10.1007/978-3-642-24085-0_69

M3 - Paper

SP - 680

EP - 690

ER -