Human eyes detection using hybrid neural method

In this paper, a novel hybrid neural method for human eyes detection is proposed. Firstly, the input facial image is preprocessed for normalization. Then, the candidates of eyes regions are detected based on a modified version of radial basis function (RBF) neural networks. Finally, a hierarchical knowledge-based approach is presented to validate these candidates and to locate the positions of both eyes. Our method has been tested on a database of facial images and achieves a good performance.