Color and face image analysis
Color
analysis is an important generic research area of machine vision. Color information is much more appropriate than gray
level information for detecting and classifying objects and surfaces with little edginess.
New developments in color vision
research are based on physical models. Using these models in color image analysis we can get simultaneously color and
geometrical parameters, e.g., shape, texture and roughness of the analyzed scene, and information on current illumination.
Our group focuses its research on development of approaches to illumination invariant color machine vision in order to
improve the performance of current color measurement and recognition techniques and to increase their domain of
applications.
A face database, which contains color images of faces under different illuminants and camera calibration conditions as well as
skin spectral reflectance measurements of each person, is collected. Using this database, methods for color correction of
face images under different illuminants are developed.
Skin color is also modeled and methods of skin detection under varying illumination conditions are proposed .
The group is also investigating methods of face detection and recognition.
|
|
