In order to make the human-computer interaction becomes more natural

Pseudo-jacobi-Fourier Moments is used for features extraction of gestures image

and combining with Euclidean distance classifier for modeling

classify and hand gesture recognition. Use kinect sensor to obtain hand depth data flow

using depth data combined with threshold segmentation method

can effectively implement gesture segmentation. Pseudo-JacobiFourier is a kind of invariant moment which has gray

rotation and scale invariance

which is suitable for feature extraction of multi-distorted invariant image. Experiments on five kind of gestures are tested

and the average recognition rate is 95. 2 %

experimental result demonstrates that the method has high recognition rate.