| Project Name : |
Integrating Multiple Visual Biometic for Human Recognition
|
| Objective : |
The objective of this project is to develop a new video based recognition system
to recognize non-cooperative individuals who appear in camera with mostly side
view or half profile view. It has been found to be difficult to recognize a person in
practical surveillance environment when the face information is incomplete.
Although face reconstruction technique can help to transform an incomplete face
into a frontal face for recognition, it can hardly work when only 30%~40% face
appears in the surveillance camera as usually happens when a non-cooperative
individual walking though. For optimal performance, a video surveillance system
should use as much information as possible from the observations. A fusion
system, which combines face and other visual information, is a potential
approach to accomplish the task of human recognition. The outcome of this
project is expected to help enhancing the performance on video-based
surveillance substantially. We expect the proposed fusion system is more
reliable than the recognition system with single measure.
|
| Project Duration : |
12 months |
Introduction