Improved RHT algorithm for intelligent recognition of key features of face

To improve the performance and precision of ordinary generalized random Hough transform (RHT) algorithm, especially in the application of key features recognizing of human face, three orientating degrees are introduced, which are X, Y, and rotate angle, to construct the factors coordination. The model can be used in the procedure of grey-scale and arbitrary objects detection, aim at 2-demension objects recognition with 3 orientating degrees, a generalized RHT algorithm is constructed with high-flexibility. Experiments prove that this model is able to demonstrate ideal ability of recognition on key arbitrary features of human faces, three orientating degrees can be detected quickly and correctly, as such speed up whole face recognition.