Human face pose estimation method and apparatus, terminal and storage medium

A face pose and terminal technology, applied in the field of image recognition, can solve problems such as the inability to detect key points at large angles, pose value errors, etc., and achieve the effects of shortening estimation time, fewer network layers, and improving efficiency

Active Publication Date: 2018-06-22
SHENZHEN INTELLIFUSION TECHNOLOGIES CO LTD
View PDF11 Cites 21 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The regression-based method mainly relies on the key points and the 3D face model, but it is still unable to detect the key points

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
  • Human face pose estimation method and apparatus, terminal and storage medium
  • Human face pose estimation method and apparatus, terminal and storage medium
  • Human face pose estimation method and apparatus, terminal and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0058] figure 1 It is a flow chart of the face pose estimation method provided by Embodiment 1 of the present invention. The described face pose estimation method is applied to a terminal.

[0059] In this embodiment, the face pose estimation method can be applied to an intelligent terminal with a camera or camera function, and the terminal is not limited to a personal computer, a smart phone, a tablet computer, a desktop or an all-in-one machine etc.

[0060] The face pose estimation method can also be applied in a hardware environment composed of a terminal and a server connected to the terminal through a network. Networks include, but are not limited to: Wide Area Networks, Metropolitan Area Networks, or Local Area Networks. The face pose estimation method in the embodiment of the present invention may be executed by a server, may also be executed by a terminal, and may also be executed jointly by the server and the terminal.

[0061] For example, for a terminal that ne...

Embodiment 2

[0087] figure 2It is a flow chart of the residual neural network training method provided by Embodiment 2 of the present invention. The residual neural network training method specifically includes the following steps. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted.

[0088] 201: Construct sample set

[0089] In this preferred embodiment, multiple face images of multiple people with different poses are prepared. You can take or collect multiple face pose images of multiple people by yourself, or you can directly obtain them from the face dataset. The face datasets include: 300-W dataset (300Faces in-the-wild), AFLW dataset, AFW dataset, Helen dataset, IBUG dataset, LFPW dataset, LFW dataset, etc.

[0090] The construction sample set specifically includes:

[0091] 1) Manually mark 68 facial key points;

[0092] In this preferred embodiment, in order to obtain correct facial posture information, ...

Embodiment 3

[0111] refer to Figure 4 Shown is a functional block diagram of a preferred embodiment of the face pose estimation device of the present invention.

[0112] In some embodiments, the face pose estimation device 40 runs in the terminal 5 . The human face pose estimation device 40 may include a plurality of functional modules composed of program code segments. The program codes of each program segment in the human face pose estimation device 40 can be stored in the memory 51 of the terminal 5, and executed by the at least one processor 52 to perform (see for details figure 1 Description) Segmentation of large-resolution face images.

[0113] In this embodiment, the human face pose estimation device 40 can be divided into multiple functional modules according to the functions it performs. The functional modules may include: an input module 401, a first classification module 402, a first output module 403, a post-processing module 404, a second classification module 405 and a s...

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 human face pose estimation method. The method comprises the steps of inputting to-be-estimated human face pose images; according to a first classification model, performing coarse classification on the to-be-estimated human face pose images, thereby identifying whether the to-be-estimated human face pose images are full side face human face images; when the first classification model identifies that the to-be-estimated human face pose images are the full side face human face images, outputting the to-be-estimated human face pose images as full side faces; when the first classification model identifies that the to-be-estimated human face pose images are not the full side face human face images, performing fine classification on the to-be-estimated human face pose images according to a second classification model; and outputting human face pose values of the to-be-estimated human face pose images. The invention furthermore provides a human face pose estimation apparatus, a terminal and a storage medium. The human face pose estimation from the coarse classification to the fine classification is realized; the human face pose estimation efficiency is improved;and a relatively good human face pose estimation effect is achieved.

Description

technical field [0001] The present invention relates to the technical field of image recognition, in particular to a face pose estimation method, device, terminal and storage medium. Background technique [0002] At present, face pose estimation plays an important role in the fields of face recognition and human-computer interaction. Face pose estimation is to estimate the pose of the face in the 2D image in the 3D space. Changes in face posture will lead to loss of face information and differences, so that the similarity between the profile faces of different people is higher than the similarity between the profile face and the front face of the same person. [0003] At present, face pose estimation on RGB images generally includes three methods: classification-based methods, face appearance-based methods, and regression-based methods. The method based on classification is to divide the faces into different categories according to a certain interval of face angles. The m...

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): G06K9/00G06K9/62G06N3/08
CPCG06N3/08G06V40/172G06F18/24
Inventor 陈淑华牟永强
Owner SHENZHEN INTELLIFUSION TECHNOLOGIES CO LTD
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