Face key-point detection method based on context and structure modeling

Abstract

The invention discloses a face key-point detection method based on context and structure modeling. The method is used to predict positions of key points of a face in a face image in a case where the face image is given, and specifically comprises the steps of: acquiring a face image data set used for face key-point detection, defining an algorithm target, and establishing a prediction network anda context network at the same time; carrying out context modeling on predicted key points and real key points; carrying out structure modeling on the predicted key points and the real key points; training a prediction model of key-point detection; and using the prediction model to detect the positions of the key points of the face in the face image. The method is suitable for use in the face key-point detection problem under a natural scene, and has superior effect and robustness in facing various complicated cases.

Description

technical field [0001] The invention belongs to the field of computer vision, in particular to a human face key point detection method based on context and structure modeling. Background technique [0002] Face key point detection is an important research problem in computer vision. Its main purpose is to analyze and interpret faces in images and videos to obtain rich information. At present, this technology has been widely used in intelligent video surveillance, intelligent transportation and commercial face recognition systems. [0003] In natural scenes, due to factors such as face pose, lighting conditions, and occlusion, face key point detection tasks still face huge challenges in terms of feature extraction, structural information modeling, and contextual information analysis. On this task, previous research work can be roughly divided into two categories: 1) methods based on image region classification and 2) methods based on point regression. Contents of the inven...

AI Technical Summary

Problems solved by technology

[0003] In natural scenes, due to factors such as face pose, lighting conditions, and occlusion, face key point detection tasks still face huge challenges in feature extraction, structural information modeling, and contextual information analysis.

Images