Human face key point detection method, system and device based on semantic alignment

Abstract

The invention belongs to the field of face recognition and particularly relates to a face key point detection method based on semantic alignment, system and device, The objective of the invention is to improve the accuracy of face key point detection. The method comprises the following steps of: obtaining a basically converged face key point detection network by a traditional method; adopting theconstructed training sample comprising the face image sample marked with the key points and the standard Gaussian response graph taking the positions of the key points as the center, and using the probability model containing the hidden variables as the target of maximum likelihood estimation to optimize the face key point detection network; And predicting the coordinates of the face key points through the finally optimized face key point detection network. According to the method, the problem of training oscillation caused by labeling randomness is effectively solved in the network training process, and the accuracy of face key point detection is improved.

Description

Technical field [0001] The invention belongs to the field of face recognition, and specifically relates to a method, system and device for detecting key points of a face based on semantic alignment. Background technique [0002] Face key points occupies a very important position in computer vision based on face, pattern recognition applications, such as video surveillance and identity recognition systems. For most face applications, it is first necessary to accurately detect the key points of the face. [0003] In recent years, the mainstream face key point detection methods are mainly divided into two categories, one is the traditional method. One is the method based on convolutional neural network. Traditional methods directly return model parameters through manual image features. The representative method is cascade regression, and the fitting process can be summarized as the following formula: [0004] p k+1 = P k +Reg k (Fea(I,p k )) [0005] In the kth iteration, pass Reg k ...

AI Technical Summary

Problems solved by technology

However, there are a large number of weak semantic points in the key points of the face. These points are usually only required to be evenly distributed on the specified edge, such as the contour of the face, the eye socket, the bridge of the nose, etc., and there is no strict semantic position

Due to the low recognition of texture information around these weak semantic points, random errors inevitably exist in the manual labeling results, which leads to semantic inconsistencies in the labels of different samples.

Therefore, using manual calibration points to directly train the model will lead to a large number of invalid errors in the training process, so that the network fitting ability cannot be concentrated where it is really needed

So far, the impact of the randomness of the labeling of key points on model training has not been paid attention to

Images