Summary
International Technical Conference on Circuits/Systems, Computers and Communications
2016
Session Number:P1
Session:
Number:P1-12
3D Alignment of Face in a Single Image
Eungyeol Song, Kyungjae Lee, Sangyoun Lee ,
pp.875-877
Publication Date:2016/7/10
Online ISSN:2188-5079
DOI:10.34385/proc.61.P1-12
PDF download (1.1MB)
Summary:
The objective of this thesis paper is to suggest a method to find a face and align 3 dimensional model landmark through 67 landmarks that describe characteristics of a face from a single image. Furthermore, using Z-value from the deformable base, a 2 dimensional image can be converted into a 3 dimensional image. The suggested algorithm is the AAM algorithm, which can detect faces precisely, and the Procruest Analysis, which, among other 3 dimensional model lanmarkds, aligns with the most precision. Using the aligned point cloud, ratation, translation, and scale matrix are calculated. A 2D face is converted to a 3D fact using this coefficient.