Method and apparatus for positioning face key points

Abstract

The invention discloses a method and apparatus for positioning face key points. The method includes the following steps: conducting rough positioning through a multi-task convolutional neural network, determining substantial positions of the face key points; extracting local regions in the peripheries of the key points, fusing the local regions extracted from the peripheries of the key points through the global cascade convolutional neural network, performing cascade positioning; finally, independently training the convolutional neural network at each key point and performing precise and fine positioning. According to the invention, the neural network has fewer total numbers and has excellent positioning effects.

Description

technical field [0001] The invention relates to digital image processing, computer vision and other technical fields, in particular to a face key point positioning method and device based on a global convolutional neural network. Background technique [0002] Key point positioning (detection) is an important problem in computer vision, which refers to locating some key parts of the face with semantic structure information, such as eyes, nose and mouth, etc., which is an important part of supervised face alignment. A step of. Key point positioning also has many practical applications, such as face recognition, facial expression analysis, and human-computer interaction applications. Due to changes in head pose, facial expression, and lighting, keypoint localization remains a very challenging problem. The traditional key point detection algorithm based on convolutional neural network first uses an overall neural network to locate key points, and then isolates each key point s...

AI Technical Summary

Problems solved by technology

Although the traditional method has high positioning accuracy, it ignores the global information of the face shape, is not robust to illumination changes, occlusion, etc., and has a high time complexity, making it difficult to apply accurate key point positioning in large-scale scenes.

Images