An Improved Fuzzy C Means and Kathunen-Loeve Transform Method for Face Detection
Ming Hu, Qiang Zhang, Zhiping Wang, Gaochong Wu · 2008
An new approach based on Kathumem-Loeve(K-L) transform and improved fuzzy C means algorithm for human face detection is proposed. In the method, the image space is partitioned to primary space and subordinate space by K-L transform, by which the human face can be detected from complex scene coarsely, but how to obtain a reasonable threshold become a key problem. Generally, most scholars adopt the value of experience to distinguish face image. In this paper, in order to estimate human face, an improved fuzzy C means method is used to classify the Euclidean distance calculated by K-L transform. The result of the experiment shows that this approach can solve the problem of human face detection in scale, location, orientation and pose from complex scene.