Key points are not available for this paper at this time.
For static gestures image classification problems, this paper proposed a kind of image recognition method based on the unsupervised feature learning, mainly using the sparse auto-encoding neural network model to extract image edge feature which is belong to unsupervised feature learning, then let the edge feature as the input of the classifier, finally sort and identify the gesture image which having depth information by the trained classifier.
Ma et al. (Mon,) studied this question.