Supercharge Your Innovation With Domain-Expert AI Agents!

Image beauty processing method, system, storage medium and device based on retinaface algorithm

A processing method and image technology, applied in the field of image processing, can solve the problem of low speed of face key point detection

Active Publication Date: 2022-08-09
江西中业智能科技有限公司
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Based on this, the object of the present invention is to provide an image beauty treatment method, system, storage medium and equipment based on the Retinaface algorithm, which is used to solve the problem of low detection speed of human face key points in the prior art

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
  • Image beauty processing method, system, storage medium and device based on retinaface algorithm
  • Image beauty processing method, system, storage medium and device based on retinaface algorithm
  • Image beauty processing method, system, storage medium and device based on retinaface algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] see figure 1 , which shows the image beautification processing method based on the Retinaface algorithm in the first embodiment of the present invention, and the method includes steps S101 to S103:

[0044] S101. Obtain original image data, extract multiple data features of the original image data through a backbone network in the algorithm, and synthesize the multiple data features into a feature pyramid.

[0045] Specifically, multiple data features of the original image data are extracted through the backbone network in the Retinaface algorithm. The backbone network includes multiple sub-networks, and each sub-network extracts data features of different dimensions. The sub-networks include classification sub-networks, and face frame detection Sub-network and face key point sub-network.

[0046] S102. Perform feature detection on the data features after synthesizing the feature pyramid through each sub-network to obtain a plurality of sub-networks after training, com...

Embodiment 2

[0050] Please check figure 2 , which is an image beautifying processing method based on the Retinaface algorithm in the second embodiment of the present invention, and the method includes steps S201 to S203:

[0051] S201. Obtain original image data, extract multiple data features of the original image data through a backbone network in the algorithm, and synthesize multiple data features into a feature pyramid.

[0052] Specifically, the backbone network includes a plurality of sub-networks, each sub-network extracts data features of different dimensions, and the sub-networks include a classification sub-network, a face frame detection sub-network and a face key point sub-network.

[0053] The training data picture and coordinates are obtained from the original image data, scaled to 640*640 pixels, and input into the backbone network. Specifically, MobilieNetV3 (0.25 times) is used as the backbone network of Retinaface. It is an example but not a limitation. In other embodim...

Embodiment 3

[0090] see Figure 5 , which is an image beautifying processing system based on the Retinaface algorithm in the third embodiment of the present invention, and the system includes:

[0091] an acquisition module for acquiring original image data, extracting multiple data features of the original image data through the backbone network in the Retinaface algorithm, and synthesizing a plurality of the data features into a feature pyramid, and the backbone network includes multiple sub-networks, Each of the sub-networks extracts data features of different dimensions, and the sub-networks include a classification sub-network, a face frame detection sub-network and a face key point sub-network;

[0092] The training module is used to perform feature detection on the data features after synthesizing the feature pyramid through each of the sub-networks to obtain a plurality of sub-networks after training, and combine a plurality of sub-networks after training to obtain the backbone net...

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 present invention provides an image beautifying processing method, system, storage medium and device based on Retinaface algorithm. The method includes: acquiring original image data, extracting multiple data features of the original image data, synthesizing the data features into a feature pyramid, and then passing through each The sub-network performs feature detection to obtain multiple trained sub-networks, merges the trained sub-networks to obtain the trained backbone network, and establishes a target detection model; The face data obtains the coordinates of multiple face key points through the target detection model, and obtains the beautified image through the face key points. The above-mentioned image beauty processing method, system, storage medium and device based on the Retinaface algorithm can directly perform key point detection through the trained model, improve the inspection speed, and solve the problem of low detection speed of face key points in the prior art. question.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an image beautifying processing method, system, storage medium and device based on the Retinaface algorithm. Background technique [0002] With the improvement of computing power of mobile devices and the development of neural network model optimization, it is possible to perform face beautification and face-lifting locally on the mobile terminal in real time. [0003] Beauty refers to the use of image processing technology to perform beautification of portraits in images or videos to better meet the aesthetic needs of users. In the field of live broadcast, it often involves using the mobile phone camera to take a selfie or shoot other people's scenes. In these scenes, real-time beauty is widely required, including functions such as filters, face-lifting, and big eyes. Among them, face-lifting and big eyes rely on the key to the face. point detection. Retinaface algori...

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): G06T5/00
CPCG06T2207/10004G06T2207/20081G06T2207/20084G06T2207/30201G06T5/77
Inventor 史绍阳李介刘丹张恒星高园岗杨忠
Owner 江西中业智能科技有限公司
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com