Projects per year
Abstract
The Iterative Closest Point (ICP) algorithm can accurately register and match 3D rigid surfaces. However, when it comes to non-rigid or semi-rigid surfaces such as the human face, the performance of ICP drops significantly. In this paper, we extend the ICP algorithm and propose a Semi-Rigid ICP algorithm which can match and register semi-rigid surfaces. We compare the performance of SRICP and ICP algorithms in a challenging scenario whereby 3D faces under exaggerated facial expressions are matched to 3D faces under neutral expression for recognition. Our results show that the proposed
SRICP algorithm performs significantly better than the original ICP algorithm.
SRICP algorithm performs significantly better than the original ICP algorithm.
Original language | English |
---|---|
Title of host publication | Proceedings Image and Vision Computing, New Zealand 2006 |
Editors | Patrice Delmas, John Morris |
Place of Publication | New Zealand |
Publisher | Image & Vision Computing, New Zealand |
Pages | 257-262 |
ISBN (Print) | 9780473117924 |
Publication status | Published - 2006 |
Event | SRICP: An Algorithm for Matching Semi-rigid 3D Surfaces - Great Barrier Island, N.Z. Duration: 1 Jan 2006 → … |
Conference
Conference | SRICP: An Algorithm for Matching Semi-rigid 3D Surfaces |
---|---|
Period | 1/01/06 → … |
Fingerprint
Dive into the research topics of 'SRICP: An Algorithm for Matching Semi-rigid 3D Surfaces'. Together they form a unique fingerprint.Projects
- 1 Finished
-
DP0664228 - Unified Representations of Multimodal Biometrics for Robust Authentication and Identification
1/01/06 → 31/12/08
Project: Research