Landmark Localisation in 3D Face Data

TY - GEN

T1 - Landmark Localisation in 3D Face Data

AU - Romero, Marcelo

AU - Pears, Nick

PY - 2009

Y1 - 2009

N2 - A comparison of several approaches that use graph matching and cascade filtering for landmark localisation in 3D face data is presented. For the first method, we apply the structural graph matching algorithm "relaxation by elimination" using a simple "distance to local plane" node property and a "Euclidean distance" arc property. After the graph matching process has eliminated unlikely candidates, the most likely triplet is selected, by exhaustive search, as the minimum Mahalanobis distance over a six dimensional space, corresponding to three node variables and three arc variables. A second method uses state-of-the-art pose-invariant feature descriptors embedded into a cascade filter to localise the nose tip. After that, local graph matching is applied to localise the inner eye corners. We evaluate our systems by computing root mean square errors of estimated landmark locations against ground truth landmark localisations within the 3D Face Recognition Grand Challenge database. Our best system, which uses a novel pose-invariant shape descriptor, scores 99.77% successful localisation of the nose and 96.82% successful localisation of the eyes.

AB - A comparison of several approaches that use graph matching and cascade filtering for landmark localisation in 3D face data is presented. For the first method, we apply the structural graph matching algorithm "relaxation by elimination" using a simple "distance to local plane" node property and a "Euclidean distance" arc property. After the graph matching process has eliminated unlikely candidates, the most likely triplet is selected, by exhaustive search, as the minimum Mahalanobis distance over a six dimensional space, corresponding to three node variables and three arc variables. A second method uses state-of-the-art pose-invariant feature descriptors embedded into a cascade filter to localise the nose tip. After that, local graph matching is applied to localise the inner eye corners. We evaluate our systems by computing root mean square errors of estimated landmark locations against ground truth landmark localisations within the 3D Face Recognition Grand Challenge database. Our best system, which uses a novel pose-invariant shape descriptor, scores 99.77% successful localisation of the nose and 96.82% successful localisation of the eyes.

KW - 3D feature descriptors

KW - facial landmark localisation

KW - cascade filter

KW - relaxation by elimination

KW - SSR histograms

KW - RECOGNITION

UR - http://www.scopus.com/inward/record.url?scp=72349095000&partnerID=8YFLogxK

M3 - Conference contribution

SN - 978-1-4244-4755-8

SP - 73

EP - 78

BT - AVSS: 2009 6TH IEEE INTERNATIONAL CONFERENCE ON ADVANCED VIDEO AND SIGNAL BASED SURVEILLANCE

PB - IEEE

CY - NEW YORK

T2 - 6th IEEE International Conference on Advanced Video and Signal Based Surveillance

Y2 - 2 September 2009 through 4 September 2009

ER -