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

A face value calculation method based on convolutional neural network

A convolutional neural network and face image technology, applied in computing, computer components, instruments, etc., can solve the problems of changing the grayscale information of the face image, the result of the image value is different, and the calculation of the value of the face is affected. It achieves the effects of good robustness, high discrimination, and improved speed of calculation of appearance value

Active Publication Date: 2018-04-27
CHINA JILIANG UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Two face images belonging to the same person may have different appearance results due to the influence of posture
[0006] 2. The change of illumination will change the grayscale information of the face image, so it will affect the calculation of the face value
[0007] 3. Face images may also be affected by age, occlusion, etc., which will affect the accuracy of face value calculation to varying degrees

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
  • A face value calculation method based on convolutional neural network
  • A face value calculation method based on convolutional neural network
  • A face value calculation method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] Hereinafter, preferred examples of the present invention will be described in detail with reference to the accompanying drawings.

[0035] like figure 1 Shown, the embodiment of the inventive method specifically comprises the following steps:

[0036] 1) Prepare face images of different people and perform necessary pre-processing to obtain ideal face images; specifically, since the convolutional neural network is used as a deep learning model, complex pre-processing of images is not required. Just perform scale normalization, normalize the image to 100×120 resolution, and the image format is an RGB color image.

[0037] 2) Extract the global face image block and local face image block of each face image according to the key points of the face; specifically, first use the key point detection method to locate the key point positions of the face, and the positions of the five key points are as follows: image 3 As shown, the global face image block and local image block ...

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 calculating the color value of a human face image based on a convolutional neural network. Collect face images, including those without face value tags and those with face value tags; perform pre-processing, obtain key points of faces and extract global and local face image blocks; pre-train convolutional neural network and then Perform fine-tuning, extract the depth feature of the face, extract the shape feature, and combine the two as the face value feature; input the face value feature into the classifier to train the face value classifier; perform the above steps on the tested face image in order to obtain The face value feature is calculated by using the face value classifier to calculate the face value feature of the tested face image. The present invention uses the convolutional neural network to extract the depth features of the global face and partial face images, and combines the face shape features to overcome the uncertainty of face value calculation in complex situations, has high robustness, and has good engineering applications. Effect.

Description

technical field [0001] The present invention relates to an image processing method, in particular to a method for calculating the appearance value of a human face image based on a convolutional neural network in the technical field of computer vision recognition. Background technique [0002] Face value indicates the handsome or beautiful value of a character's appearance, and is used to evaluate the character's appearance. "Xiangyou" social software initiates the appearance ranking list, and calculates the user's appearance level according to the number of "no feeling" and "likes" received by the user, for example: a user receives 100 comments from other users If there are 70 'likes' in the feedback, the user's appearance value is 0.7 (0.0-1.0). The higher the face value, the better the appearance, and the lower the face, the less good-looking. The calculation of face value mainly refers to whether the proportions of the facial features are coordinated, and the incisive c...

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 Patents(China)
IPC IPC(8): G06K9/00
Inventor 郑恩辉陈良仁富雅琼陈乐
Owner CHINA JILIANG UNIV
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