Site Network: Personal | CASIA

Research Statements

My research focuses on Face Recognition,Face Alignment,Face Cloning, photo-realistic 3D face models,3D Finger Interaction, and 3D capturing.


Face Recognition

Face Recognition

we present parallel feature selection mothed to select complementary features from extended Gabor wavelet. A very efficient face regonition engine is constructed for FRGC test and real system. It is real-time in embedded platform.

 


Face Alignment

Face Alignment

In this project, we present a fast and robust face alignment and expression tracking system. The system can locate face key points, track the head pose and facial expression with very high stability in real time speed.

 

 


Face Cloning

Face Cloning

We are developing techniques for cloning facial expressions displayed by one person to face models representing other faces. A system capable of doing this in real-time could then be used to manipulate faces during live face-to-face interactions such that the influences of apparent appearance versus social expectation can be studied.

 

 


photo-realistic 3D face models

photo-realistic 3D face models

A novel method was developed for creating photo-realistic 3D face models from 2D images of a subject. A scattered data interpolation algorithm was used to deform an exemplar mesh to fit individual face geometries.

 

 


3D Face Registration


3D Face Registration

In this paper, we present a system to realize 3D modeling in real time. A person just needs to appear in front of the sensor and the system will detect the face region and obtain the depth data. While the user rotating his head around, the system tracks the pose and adds the new data to the face model incrementally.

 

 


3D Finger Interaction


3D Finger Interaction

We present a novel solution to the problem of recovering and tracking the 3D position, orientation and full articulation of a human hand from markerless visual observations obtained by a Kinect sensor.The proposed method does not require special markers and/or a complex image acquisition setup. Being model based, it provides continuous solutions to the problem of tracking hand articulations.A prototype CPU-based implementation of the proposed method demonstrate that accurate and robust 3D tracking of hand articulations can be achieved in near real-time.

 

 


Computational Plenoptic 3D Imaging


Computational Plenoptic 3D Imaging

The plenoptic function is a ray-based model for light that includes the color spectrum as well as spatial, temporal, and directional variation. In this project, we review approaches that optically encode dimensions of the plenpotic function beyond those captured by traditional photography and reconstruct the recorded information computationally.

 

 

 

Last Updated: January 5, 2013