Face recognition using image-set or video sequence as input tends to be more robust since image-set or video sequence provides much more information than single snapshot about the ...
This paper proposes a novel approach to motion capture
from multiple, synchronized video streams, specifically
aimed at recording dense and accurate models of the structure
and ...
Yasutaka Furukawa (University of Washington), Jean...
This paper presents Support Vector Machine (SVM) with local summation kernel for robust face recognition. In recent years, the effectiveness of SVM and local features is reported....
We propose an algorithm for extracting facial features robustly from images for face recognition under large pose variation. Rectangular facial features are retrieved via the by-p...
Evaluations of the state-of-the-art of both academic face recognition algorithms and commercial systems have shown that recognition performance of most current technologies degrad...