Abstract | ||
---|---|---|
This paper investigates head pose estimation problem which is considered as front-end preprocessing for improving multi-view human face recognition. We propose a computational model for perceiving head poses based on the Non-negative Multi-way Factorization (NMWF). The model consists of three components: the tensor representation for multi-view faces, feature selection and head pose perception. To find the facial representation basis, the NMWF algorithm is applied to the training data of facial images. The face tensor includes three factors of facial images, poses, and people. The former two factors are used to construct the computational model for pose estimation. The discriminative measure for perceiving the head pose is defined by the similarity between tensor basis and the representation of testing facial image which is projection of faces on the subspace spanned by the basis "TensorFaces". Computer simulation results show that the proposed model achieved satisfactory accuracy for estimating head poses of facial images in the CAS-PEAL face database. |
Year | DOI | Venue |
---|---|---|
2007 | 10.1007/978-3-540-69158-7_86 | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
Keywords | Field | DocType |
tensor representation,facial representation basis,multi-view human face recognition,perceiving head,cas-peal face database,face tensor,tensor basis,facial image,computational model,tensor factorization,head pose estimation,computer model,face recognition,pose estimation,front end,feature selection,computer simulation | Tensor,Feature selection,Computer science,Pose,Artificial intelligence,Articulated body pose estimation,Discriminative model,Computer vision,Facial recognition system,Subspace topology,Pattern recognition,3D pose estimation,Machine learning | Conference |
Volume | Issue | ISSN |
4984 LNCS | PART 1 | 16113349 |
Citations | PageRank | References |
1 | 0.35 | 22 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wenlu Yang | 1 | 28 | 7.81 |
Liqing Zhang | 2 | 2713 | 181.40 |
Wenjun Zhu | 3 | 18 | 4.98 |