Recognition of face direction

From the observations of flesh-colored pixel distributions on UV plane in the LUV color space, an ellipse successfully extracted a face region. A proposed method using ellipse extraction and compensation with Soble boundary was an extension of the method proposed by Adachi et al. Extracted rotate face images by 15 degrees interval were used to make an automatic face direction detecting system. The face direction was estimated through two processes. The first process was a projection into an eigenspace, and the second process was an estimation of direction through analytical interpolation. At first, a seven-dimensional eigenspace was obtained from the principal component analysis of thirteen face images. Then, from the distances in a eigenspace, the face direction was detected analytically. The accuracy of detection of face directions was adequate.