In:
Jurnal Teknologi, Penerbit UTM Press, Vol. 75, No. 2 ( 2015-07-13)
Abstract:
This paper analyzes the techniques that can be used to perform point cloud data registration for a human face. We found that there is a limitation in full scale viewing on the input data taken from 3D camera which is only represented the front face of a man as the point of view of a camera. This has caused a hole on the surface that is not filled with the point cloud data. This research is done by mapping the retrieved point cloud to the surface of the face template of the human head. By using Coherent Point Drift (CPD) algorithm which is one of the non-rigid registration techniques, the analysis has been done and it shows that the mapping of points for a three-dimensional (3D) face is not done properly where there are some surfaces work well and certain points spread into the wrong area. Consequently, it has resulted in registration failure occurrences due to the concentration of the points which is focusing on the face only.
Type of Medium:
Online Resource
ISSN:
2180-3722
,
0127-9696
DOI:
10.11113/jt.v75.4987
Language:
Unknown
Publisher:
Penerbit UTM Press
Publication Date:
2015
detail.hit.zdb_id:
2780014-3
