Abstract:

This paper introduces the gender recognition method based on neural networks for the difficulty in extracting the human secondary sex characteristics,and the low recognition rate and robustness.We get the gender recognition classifier based on human face images.The face image is processed by the Gauss filter,and then the normalization of preprocessed image is used to train the BP neural network to get the classifier.The gender recognition method is compared with the traditional ones.The experimental results show that the extraction of the human secondary sex characteristics can be automatically done by our method,and the recognition rate,robustness and the fault tolerance are also improved.

Key words: gender recognition;neural networks;gaussian filter;robustness