Algorithm Of Gesture Recognition Descriptors Based On Fourier

The use of hand gestures has become an important part of human computer interaction in recent years. The ability for computer to visually recognize hand gestures is essential for future human computer interaction. This thesis presents a vision-based hand gestures recognition algorithm from points of pre-processing, feature extraction and recognition of hand gestures image.Preprocessing hand gesture image is composed of three parts, image enhancement, image segmentation and morphological image processing. The thesis firstly performs image smoothing and sharpening, then gets the binary version of the images by the means of a gray level threshold algorithm. After that,to get a better binary image the system takes the operation of morphological filtering.In the part of feature extraction, Focusing on the problem of low recognition rate and large noise interference in the current hand gestures recognition, a new algorithm based on two-dimensional polar Fourier transform is presented in this thesis. The descriptor in this algorithm is invariant to general transformation including translation,scale and rotation transformation. Compared with traditional one-dimensional descriptor, this two-dimensional descriptor uses the boundary information of a hand gestures image and also extracts the information inside it. Therefore the descriptor is strongly applicable and highly robust.In the part of feature recognition,a method based on BP neural networks is used in this thesis. First, the BP network is trained with a large number of sample images. After the network is trained successfully, the system can recognize hand gesture images, and the validity of identification can be determined according to the recognition result.The experimental results show that the algorithm based on two-dimensional polar Fourier transform and BP neural networks is very efficient and the recognition rate is 94%.