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

A method for generating a certificate map Trimap graph by using a full convolutional neural network

A convolutional neural network and ID photo technology, applied in the field of using full convolutional neural network to generate ID photo Trimap, can solve problems such as inability to obtain effects, achieve convenience and effect improvement, high robustness, and good noise resistance Effect

Active Publication Date: 2018-12-11
SOUTH CHINA UNIV OF TECH
View PDF3 Cites 13 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method does not achieve ideal results when the background and foreground contrast is low or the unknown area is large.

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 method for generating a certificate map Trimap graph by using a full convolutional neural network
  • A method for generating a certificate map Trimap graph by using a full convolutional neural network
  • A method for generating a certificate map Trimap graph by using a full convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0041] The present invention will be further described below in conjunction with specific examples.

[0042] The method provided in this example using the full convolutional neural network to generate the Trimap image of the certificate is as follows:

[0043] 1) Data set preparation

[0044] The training first needs to establish a ID photo dataset containing the broad shoulder area of ​​the human head portrait, such as figure 1 shown. The source of the data set is a batch of ID photos that need to be collected for laboratory projects. After collection, they are screened to ensure that they cover men, women, the elderly, adults, children, long hair, short hair, straight hair, curly hair, and people of various colors. Clothes, gray and blue backgrounds for different ID photos, and weed out those that are unevenly exposed and out of focus. All images are cropped to the same size based on the position of the portrait to ensure the usability of the portrait. The dataset contai...

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 certificate Trimap diagram by using a full convolutional neural network. The method comprises the following steps: 1) carrying out data input; 2) training a full convolutional neural network model; 3) segmenting the input image using the trained model. The method mainly solves the problem that a network model is constructed through a self-established certificate photograph data set including a human head image and a shoulder area, and the network model is trained by using the data set. After the training, the network model will be able to segment the human head image including shoulder region with high accuracy. The method has the advantages of high accuracy, good noise immunity, simple use, high efficiency, high speed and the like. Compared with a conventional method, the method realizes the Trimap image segmentation processing for the certificate photo for the first time, and provides good input for matting. At the same time, the method of the invention achieves better effect in the recognition and segmentation of shoulders and clothes.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a method for generating a Trimap image of a certificate photo using a fully convolutional neural network. Background technique [0002] The ID photo is a photo that proves the identity of the holder in various types of certificates. People need to apply for various certificates in their daily life, and the photos taken for making these certificates also need to be processed before they can be used. A very common way to process ID photos is to replace the background color of photos. The core technologies involved include image segmentation and matting. Most of the matting algorithms with ideal effects depend on the input of Trimap images, and the quality of Trimap images also directly affects the matting results. At present, there are very few studies on the automatic generation of Trimap images. Manually marking images consumes manpower and time. It is not small, so the...

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): G06T7/194G06T7/11
CPCG06T7/11G06T7/194G06T2207/20081G06T2207/20084G06T2207/30196
Inventor 邹歆仪李桂清
Owner SOUTH CHINA UNIV OF TECH
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