Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Head attitude estimation method based on depth learning

A head posture, deep learning technology, applied in the field of computer vision, can solve problems such as weak versatility, and achieve the effect of solving weak prediction robustness

Inactive Publication Date: 2018-02-23
SEETATECH BEIJING TECH CO LTD
View PDF12 Cites 39 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this type of method requires a specific input device, so its versatility is not strong

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
  • Head attitude estimation method based on depth learning
  • Head attitude estimation method based on depth learning
  • Head attitude estimation method based on depth learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0022] A head pose estimation method based on deep learning, the specific steps are:

[0023] (1) Data preparation:

[0024] The UmdFaces dataset is selected as the training sample; the data angles in UmdFaces change continuously, and there are 367,920 face photos and 8,501 different characters, with a large amount of data and sufficient samples. The schematic diagram of human head rotation posture in three-dimensional coordinates is as follows: figure 1 shown.

[0025] Due to the lack of large-angle data, especially in the yaw and pitch directions, the proportion of large-angle images is less than 5%. During the test of the initial training model, it was found that the maximum detection result in the yaw direction can only reach 53°, and the pitch The orientation can only reach 30°. In response to this problem, the data augmentat...

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 head attitude estimation method based on depth learning. The method comprises a step of acquiring an image data set for training and carrying out information labeling of a face head deflection angle on image data, a step of carrying out sample expanding and preprocessing on the data set and cutting out a face part, a step of zooming all face images which are subjected topreprocessing into resolutions of 90*90 pixels, a step of taking the above data set as a training sample and carrying out network training by using a depth network TinyPoseNet, and a step of extracting a trained TinyPoseNet network model, obtaining a cut face image of an image which needs to be tested according to the above steps, cutting an area of 80*80 pixels in the middle of the image, carrying out forward calculation of the TinyPoseNet network model and thus estimating an angle of face head posture deflection in the tested image. The method has the advantages of an ultra-small calculationamount, strong robustness, high accuracy, fast calculation, simple operation and strong universality.

Description

technical field [0001] The invention relates to an estimation method, in particular to a head posture estimation method based on deep learning, which belongs to the technical field of computer vision. Background technique [0002] Head posture is an inherent attribute of human beings, which has important application value in the fields of human emotion recognition, fatigue state monitoring, and living body verification. In general, head pose estimation is based on the three directions of pitch, yaw, and roll. It is assumed that the head pose activity is regarded as a rigid body motion, with the tip of the nose as the origin, the horizontal direction as the x-axis, the vertical direction as the y-axis, and the z-axis perpendicular to The plane formed by the x and y axes, then the angle of clockwise rotation around the x, y, and z axes is defined as the offset angle of the head posture in the pitch, yaw, and roll directions. Due to the influence of illumination, occlusion, re...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V40/172G06N3/045G06F18/214G06F18/24
Inventor 李珊如刘昕袁基睿山世光
Owner SEETATECH BEIJING TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com