Method for making human face posture estimation utilizing dimension reduction method

Abstract

This invention discloses a method for estimating man-face gestures by a dimensionality reduction method including the following steps: 1, preprocessing man-face image trained samples of different gestures, 2, carrying out PCA process to processed data, 3, initializing limited nerve network of a Boltzmann machine, 4, pre-training the limited Boltzmann machine nerve network with the data processed by PCA, 5, adjusting the limited Boltzmqnn nerve network parameters, 6, identifying gestures of new man-face images, which reduces the error rate even further compared with the present technology and reduces its dimensions greatly.

Description

technical field [0001] The invention relates to a method in the technical field of image recognition, in particular to a method for estimating face posture by combining principal component analysis (PCA) and a dimensionality reduction method of a restricted Boltzmann machine neural network. Background technique [0002] At present, the research focus of face recognition has been extended from 2D to 3D, and the core technology in 3D face recognition is how to estimate the 3D pose of the face based on the 2D face image. Pose estimation is essentially a classification problem, that is, to judge which pose a face in a face image belongs to. However, face images are typical high-dimensional data, and general classification methods cannot be directly applied to pose estimation. Therefore, it is necessary to reduce the dimensionality of high-dimensional face data first, and then perform pose estimation on the reduced-dimensional data. [0003] Through the retrieval of prior art do...

AI Technical Summary

Problems solved by technology

However, the training time of this method increases sharply with the number of training samples and the number of dimensions.

However, in practical applications, the face image is disturbed by the outside world during the generation process and contains a lot of redundant information and noise. These redundant information and noise will not only increase the calculation time, but also affect the recognition accuracy in the later stage. Therefore, It is necessary to remove these redundant information and noise before actually using the neural network to reduce the dimensionality of these data. In addition, the restricted Boltzmann machine neural network dimensionality reduction method also has defects such as slow speed and low precision.

Images