Recognizing Face Features Using Points of Active Contour, PHP – Image Processing and GD Grafic Extension

Face recognition is one of the most spectacular and difficult biometric recognition methods because of the influence of the external factors like: distance, illumination, changes of person´s appearances, glasses and other factors. The active contour method can be used to determine face features in a picture with following amendments: the image is filtered with PHP- image processing and GD graphic extension, the energies calculated in active contour algorithm are adapted so the continuous energy is approximated with the distance between two consecutive points, the curvature energy is normalized using the same distance and the vicinity of each point are calculated until an edge is detected and that point will not ¿move¿ anywhere. Another condition imposed by our algorithm is that points are forced to avoid moving along the horizontal and vertical axis and the result is a much faster movement. Detection of the face features can continue inside of the area detected as the face contour. Farther research direction will be oriented to approximate the shapes detected with the active contour algorithm using functions so that the bitmap image is transformed in a vector image facilitating the process of recognition.