BRDF estimation for faces from a sparse dataset using a neural network

Hansen, Mark F; Atkinson, Gary; Smith, Melvyn

Abstract

We present a novel �ve source near-infrared photometric
stereo 3D face capture device. The accuracy of the system is demonstrated by a comparison with ground truth from a commercial 3D scanner. We also use the data from the �ve captured images to model the Bi-directional Reectance Distribution Function (BRDF) in order to synthesise images from novel lighting directions. A comparison of these synthetic images created from modelling the BRDF using a three layer neural network, a linear interpolation method and the Lambertian model is given, which shows that the neural network proves to be the most photo-realistic.

• Presentation Conference Type: Conference Paper (unpublished)
• Conference Name: Computer Analysis of Images and Patterns, CAIP 2013
• Start Date: Aug 27, 2013
• End Date: Aug 29, 2013
• Publication Date: Aug 27, 2013
• Publicly Available Date: Jun 7, 2019
• Peer Reviewed: Peer Reviewed
• Keywords: BDRF, neural network
• Public URL: https://uwe-repository.worktribe.com/output/928903
• Publisher URL: http://www.cs.york.ac.uk/cvpr/caip2013/Program.php
• Additional Information: Title of Conference or Conference Proceedings : Computer Analysis of Images and Patterns, CAIP 2013