A Face Keypoint Detection Method Based on Sampling Convolution

Abstract

The invention discloses a face key point detection method based on sampling convolution, and belongs to the technical field of image detection, and the method comprises the following steps: S1, obtaining a grayscale image containing a face, and obtaining a face frame in the grayscale image through employing a face detection algorithm; s2, preparing a training set, and performing Prussian analysison all faces in an image of the training set to obtain average face key points; s3, amplifying the average face key points according to the face frame size obtained in the step S1 to obtain initial face key points; s4, updating the face key points by using the network model generated by training to obtain final face key points; point convolution is carried out near a key point, and a result is continuously iterated and updated, so that the calculation speed is further increased while the precision is ensured.

Description

technical field [0001] The present invention relates to the technical field of image detection, in particular to a method for detecting human face key points based on sampling convolution. Background technique [0002] Deep learning has developed rapidly in recent years. Represented by neural networks, it has solved problems that were difficult to solve in many fields before. Face key point detection is the most important step before face alignment. In the application field based on face recognition (face recognition) technology, key point detection plays an important role in face recognition; similarly, the quality of key points directly It is related to the efficiency of the detector to identify the target. [0003] Face key point detection methods are roughly divided into three types, which are traditional methods based on ASM (Active Shape Model) and AAM (Active Appearnce Model), methods based on cascaded shape regression, and methods based on deep learning. [0004] C...

AI Technical Summary

Problems solved by technology

[0004] Currently, deep learning is the best algorithm for facial key point detection. Most of the algorithms use convolutional neural networks, and convolution operations are usually time-consuming. Some researchers have begun to use convolutional networks on small images. detection, in order to increase the calculation speed, but at the expense of accuracy

Therefore, there is currently no algorithm that can better balance the calculation speed and accuracy.

Images