A new feature space transformation algorithm in color image is presented.Basing on high dimensional imaginal geometry and biomimetics informatics and taking advantage of variations in color information

this new algorithm transforms the color image into feature space and rectifies each feature value within its neighborhood.By using this method

image feature without the effect of different light conditions can be obtained.In face detection application

unlike traditional skin tune detection method

the new method directly transforms the color image into feature space and locates the face position in feature space by face template feature matching.In this paper

the application of this new method to face detection in complex light color images has also been discussed.According to experimental result

it can be seen that the proposed method is very efficient and has significant value in application.