For the purpose of human-computer interaction(HCI),a visual approach based on gesture recognition is proposed in this paper.The technique essentially includes detection and segmentation,feature extraction,and recognition of a number of gestures which are assigned as some control commands.Each of the processing stages employs a neural network for skin-color detection,principal component analysis(PCA) as well as clustering encoding of the hand-gestures.Details of the approach and experiment results are provided.The experimental recognizing accuracy is 94 %.