Abstract. The performance of face recognition systems that use twodimensional images is dependent on consistent conditions such as lighting, pose, and facial appearance. We are dev...
We present a robust elastic and partial matching metric
for face recognition. To handle challenges such as pose, facial
expression and partial occlusion, we enable both elastic
...
An original approach to represent 2D and 3D faces using Radial Geodesic Distances (RGDs) is proposed in this work. In 3D, the RGD of a generic point of the face surface is compute...
Stefano Berretti, Alberto Del Bimbo, Pietro Pala, ...
A major goal for face recognition is to identify faces where the pose of the probe is different from the stored face. Typical feature vectors vary more with pose than with identit...
We present a novel approach to address the representation issue and the matching issue in face recognition (verification). Firstly, our approach encodes the micro-structures of t...