Abstract
3D Morphable Face Models (3DMM) have been used in pattern recognition for some time now. They have been applied as a basis for 3D face recognition, as well as in an assistive role for 2D face recognition to perform geometric and photometric normalisation of the input image, or in 2D face recognition system training. The statistical distribution underlying 3DMM is Gaussian. However, the single-Gaussian model seems at odds with reality when we consider different cohorts of data, e.g. Black and Chinese faces. Their means are clearly different. This paper introduces the Gaussian Mixture 3DMM (GM-3DMM) which models the global population as a mixture of Gaussian subpopulations, each with its own mean. The proposed GM-3DMM extends the traditional 3DMM naturally, by adopting a shared covariance structure to mitigate small sample estimation problems associated with data in high dimensional spaces. We construct a GM-3DMM, the training of which involves a multiple cohort dataset, SURREY-JNU, comprising 942 3D face scans of people with mixed backgrounds. Experiments in fitting the GM-3DMM to 2D face images to facilitate their geometric and photometric normalisation for pose and illumination invariant face recognition demonstrate the merits of the proposed mixture of Gaussians 3D face model.
Highlights
Face recognition technology has made an immense progress during the last decade, first thanks to the advances in face representation in the form of innovative features such as Local Binary Patterns (LBP) [38, 39], Local Phase Quan5 tisation (LPQ) [6, 40] and Binarised Statistical Image Features (BSIF) [29], and more recently, to the capabilities of end-to-end deep learning neural networks [51, 53]
In this paper we propose a Gaussian mixture 3D morphable face model (GM3DMM) constructed using Caucasian, Chinese and African 3D face data
We show that fitting GM-3D Morphable Face Models (3DMM) to an input 2D face image is more accurate for two reasons
Summary
Face recognition technology has made an immense progress during the last decade, first thanks to the advances in face representation in the form of innovative features such as Local Binary Patterns (LBP) [38, 39], Local Phase Quan tisation (LPQ) [6, 40] and Binarised Statistical Image Features (BSIF) [29], and more recently, to the capabilities of end-to-end deep learning neural networks [51, 53]. 35 Most of the 3D face models are of the morphable variety [2, 3] Their construction involves the use of Principal Component Analysis (PCA) to decorrelate 3D face data represented in terms of a 3D mesh of spatial coordinate samples and associated RGB surface texture values. 55 By changing the shape and texture parameters, different samples can be drawn from the morphable model. A similar list of differences could be found between, say, the African 70 ethnicity, and the above two groups This analysis suggests that a more appropriate model to construct is a mixture of Gaussians model, where each cohort of distinct ethnicity constitutes a mode in the distribution. In this paper we propose a Gaussian mixture 3D morphable face model (GM3DMM) constructed using Caucasian, Chinese and African 3D face data.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
