Human face key point detection method and apparatus

Abstract

The invention relates to the technical field of computer vision processing, provides a human face key point detection method, and aims at solving the technical problems of relatively high calculationcost and relatively low calculation resource utilization rate in human face key point detection. For the purpose, the human face key point detection method comprises the steps of identifying a human face image by utilizing a pre-constructed human face key point identification model to obtain position information of human face key points; training the human face key point identification model: performing rough key point positioning on training sample data by utilizing a distance loss function and a reverse propagation algorithm through a correction network; carrying out conversion on rough keypoints positioned by the correction network through utilizing a correction matrix to obtain corrected sample data; and carrying out regression calculation on the corrected sample data by a regressionnetwork, thereby detecting out the position information of the human face key points. Based on the above steps, the human face key point information in the human face image can be rapidly and accurately detected out.

Description

technical field [0001] The invention relates to the technical field of computer vision processing, in particular to the technical field of image processing and pattern recognition based on a non-rigid curve fitting loss function, and in particular to a face key point detection method and device. Background technique [0002] With the development of computer graphics and vision technology and the development of human-computer interaction technology, it is becoming more and more important to identify the identity information of the person being monitored in time. Identity recognition has applications in areas such as autonomous driving, human-computer interaction, smart cities, intelligent transportation, and intelligent monitoring. Among them, the identification of the person is mainly completed by the detection of the key points of the face. [0003] At present, face key point detection methods are mainly divided into two categories: regression-based face detection methods ...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above problems in the prior art, that is, in order to solve the calculation cost caused by the dependence on the number of key points in the face key point detection, and the loss function directly returns the coordinates of all key points, ignoring different key points Semantics, resulting in low utilization of computing resources, the application provides a face key point detection method and device to solve the above problems

Images