Gender Recognition with Face Images Based on PARCONE Model

In this paper, a new type of neural network model——PARCONE (Partially Connected Neural Evolutionary) was proposed, which can overcome the disadvantage that the previous neural networks can not accept more than thousands of inputs. With this new model, no feature extraction is needed before target identification and all of the pixels of a sample image can be used as the inputs of the neural network. After 300 ~ 600 generations’ evolution, the new neural network can reach a good recognition rate. With this new model, a gender recognitionwe experiment was make on 490 face images (245 females and 245 males from Color FERET database), in which include not only frontal faces but also the faces rotated from -40°~40°in the direction of horizontal. The gender recognition rate, rejection rate and error rate of the positive examples respectively achieve 95.14%, 2.16% and 2.7%. The experimental results show that the new neural model has a strong pattern recognition ability and can be applied to many other pattern recognitions whick need a large amount of input information.