Presentation 2008-06-20
3D Facial Feature Point Estimation from Noisy and Occluded 2D Measurements
Simon CLIPPINGDALE, Mahito FUJII, Nobuyuki YAGI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We discuss the simultaneous estimation, from two-dimensional measurements on a video sequence or a collection of still images of a person, of the three-dimensional positions of a number of facial feature points together with the pose of the face in each frame or image. When all feature points are visible in all frames, the SVD-based factorization of the data matrix is a well-known approach, but in the case of occlusions a partial initial solution must be extended in stages, with the result depending on the initialization and on the extension order, and requiring post-correction. Here we present a partitioned factorization algorithm that iteratively and alternately estimates pose and shape, and unlike the SVD-based approach, uses all the data uniformly.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) face / feature points / three-dimensional shape estimation / factorization / shape from motion
Paper # DE2008-24,PRMU2008-42
Date of Issue

Conference Information
Committee PRMU
Conference Date 2008/6/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) 3D Facial Feature Point Estimation from Noisy and Occluded 2D Measurements
Sub Title (in English)
Keyword(1) face
Keyword(2) feature points
Keyword(3) three-dimensional shape estimation
Keyword(4) factorization
Keyword(5) shape from motion
1st Author's Name Simon CLIPPINGDALE
1st Author's Affiliation NHK Science & Technology Research Laboratories()
2nd Author's Name Mahito FUJII
2nd Author's Affiliation NHK Science & Technology Research Laboratories
3rd Author's Name Nobuyuki YAGI
3rd Author's Affiliation NHK Science & Technology Research Laboratories
Date 2008-06-20
Paper # DE2008-24,PRMU2008-42
Volume (vol) vol.108
Number (no) 94
Page pp.pp.-
#Pages 6
Date of Issue