Model-Based multi-view face construction and recognition in videos

Authors:
Chao Wang;Yunhong Wang;Zhaoxiang Zhang;Yiding Wang
Affiliations:
School of Computer Science and Engineering, Beihang University, China;School of Computer Science and Engineering, Beihang University, China;School of Computer Science and Engineering, Beihang University, China;School of Information Engineering, North China University of Technology, China
Venue:
ICIC'12 Proceedings of the 8th international conference on Intelligent Computing Theories and Applications
Year:
2012

Citing 5
Cited 0

Active Appearance Models

IEEE Transactions on Pattern Analysis and Machine Intelligence
Kernel-Based Object Tracking

IEEE Transactions on Pattern Analysis and Machine Intelligence
An iterative image registration technique with an application to stereo vision

IJCAI'81 Proceedings of the 7th international joint conference on Artificial intelligence - Volume 2
Patch-based face recognition from video

ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
The CAS-PEAL Large-Scale Chinese Face Database and Baseline Evaluations

IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans

Quantified Score

Hi-index	0.00

Visualization

Abstract

Model-based face construction and recognition in videos is a fundamental topic in image processing and video representation, while analysis faces across multiple views is more challenging than that from a fixed view because of the severe non-linearity caused by rotation in depth, self-occlusion, self-shading and illumination. In this paper, a novel method is presented to model and recognize multi-view faces in video sequences. Firstly, we design a multi-view face model to extract the face feature points. Secondly, a hybrid tracking method integrated optical flow with mean shift is proposed to estimate the face posture. Then, by using faces' paths in different view and feature points obtained from models, a multi-view face map is synthesized by reconstruction and stitching the paths together. Finally, recognition experiments are conducted to evaluate the performance of our proposed approach.