Method for generating matching model, and face feature point tracking method

A matching model and feature point technology, applied in the field of deep learning, can solve problems such as tracking errors, weak features and learning methods, and achieve the effect of improving accuracy

Active Publication Date: 2019-09-24
XIAMEN MEITUZHIJIA TECH
View PDF15 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Due to weak features and learning methods, tracking errors usually occur for parts with less obvious features, such as key points on lips and face contours

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Method for generating matching model, and face feature point tracking method
  • Method for generating matching model, and face feature point tracking method
  • Method for generating matching model, and face feature point tracking method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided for more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0029] Traditional restricted local models consist of local experts and global shape models. The whole process is divided into model building phase and point fitting phase. The model construction stage is divided into overall shape model construction and local matching model construction. The shape model construction is to use the active shape model ASM to generate a point distribution function, concatenate the coordinates of seve...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a method for generating a matching model, and a face feature point tracking method. The matching model is suitable for performing feature point matching on a face image marked with feature points. The method for generating the matching model comprises the following steps: generating an image block taking a feature point as a center based on a face image marked with the feature point; generating an artificial label graph based on the image blocks; inputting the artificial label graph into a pre-trained convolutional neural network so as to output a prediction label graph; and training the convolutional neural network based on a loss value between the artificial label graph and the prediction label graph, and taking the trained network as a generated matching model. According to the scheme, the feature point matching accuracy can be improved, so that the face feature point tracking stability and accuracy are improved.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to a method for generating a matching model and a method for tracking facial feature points. Background technique [0002] Face feature point positioning is also called face alignment. Its goal is to locate points in the face image that can describe facial features, such as the corners of the eyes, the tip of the nose, the corners of the mouth, and the chin. At present, there are many mature methods for locating facial feature points in static images, such as supervised gradient descent method and local binarization method. However, for dynamic pictures such as the tracking and positioning of facial feature points in a video stream, since the timing of the previous and subsequent frames in the video stream is not considered, severe feature point jitter usually occurs. [0003] With the real-time filter, beauty function and augmented reality application of the beauty camera, t...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00G06N3/04G06N3/08
CPCG06N3/08G06V40/161G06V40/172G06N3/045
Inventor 王喆许清泉张伟洪炜冬曾志勇
Owner XIAMEN MEITUZHIJIA TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap