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

Face editor training, face editing and live broadcasting methods and related devices

A training method and editor technology, applied in the field of computer vision, can solve problems such as huge structure, achieve the effect of less resources, excellent performance, and simplified structure

Pending Publication Date: 2021-08-13
GUANGZHOU HUYA TECH CO LTD
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The embodiment of the present invention proposes a face editor training, face editing, live broadcast method and related devices to solve the problem of relatively large structure of tools for editing face attributes

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
  • Face editor training, face editing and live broadcasting methods and related devices
  • Face editor training, face editing and live broadcasting methods and related devices
  • Face editor training, face editing and live broadcasting methods and related devices

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] figure 2 A flow chart of a training method for a face editor provided in Embodiment 1 of the present invention. This embodiment is applicable to the situation of training a face editor for each face attribute, and the method can be performed by the face editor. Training device to perform, the training device of the face editor can be realized by software and / or hardware, can be configured in computer equipment, for example, server, workstation, personal computer, etc., specifically include the following steps:

[0052] Step 201. Using the original face data as a sample, train a face reconstructor for restoring a face.

[0053]In order to facilitate the collection of a sufficient number of data sets, image data representing faces can be collected in open source databases (such as FFHQ (Flickr Faces HighQuality)) and / or open source projects as original face data. Of course, in order to improve face reconstruction The accuracy of the face editor and the face editor in th...

Embodiment 2

[0195] Figure 8 It is a flow chart of a training method for a face editor provided by Embodiment 2 of the present invention. This embodiment is based on the foregoing embodiments and further increases the operation of the dissociation normal vector. The method specifically includes the following steps:

[0196] Step 801. Using the original face data as a sample, train a face reconstructor for restoring a face.

[0197] Step 802, training hidden vectors on the original face data.

[0198] Among them, the latent vector represents the attributes of the face.

[0199] Step 803, find the normal vector in the space where the hidden vector is located.

[0200] Among them, the normal vector represents the direction of editing each face attribute.

[0201] Step 804, for the current normal vector, set other normal vectors as reference vectors.

[0202] Step 805, dissociate the reference vector from the current normal vector as the dissociated normal vector.

[0203] In practical a...

Embodiment 3

[0217] Figure 11 A flow chart of a face editing method provided by Embodiment 1 of the present invention, this embodiment is applicable to each situation, the method can be executed by a face editing device, and the face editing device can be implemented by software and / or Or hardware implementation, which can be configured in computer equipment, such as servers, workstations, personal computers, mobile terminals (such as mobile phones, tablet computers, smart wearable devices, etc.), etc., specifically including the following steps:

[0218] Step 1101, receiving original image data.

[0219]If this embodiment is applied to user-end computer equipment such as mobile terminals and wearable devices, applications that can perform image processing can be branched in the computer equipment, such as live broadcast applications, image editing applications, camera applications, instant messaging tools, and gallery applications. etc.

[0220] For applications such as image editing a...

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 embodiment of the invention provides a face editor training method, a face editing method, a live broadcasting method and related devices, and the face editor training method comprises the steps: taking original face data as a sample, training a face reconstruction device used for restoring a face, training an implicit vector for the original face data, and representing a face attribute by the implicit vector, in the space where the implicit vector is located, searching for a normal vector which represents the direction for editing each face attribute, aiming at each face attribute, adjusting the implicit vector along the normal vector so as to generate target face data in a face reconstructor, aiming at each face attribute, and under the supervision of the paired original face data and target face data, adjusting the face reconstructor into a face editor for editing the face attribute. By independently training the face editor for the single face attribute, the face editor can keep better performance under the condition that a simple structure is applied to carry out supervised learning, the effect of editing the face attribute is controllable, and the structure of the face editor can be greatly simplified.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of computer vision, and in particular to a face editor training, face editing, live broadcast method and related devices. Background technique [0002] In scenes such as taking photos, recording videos, making short videos, and live broadcasting, use customized tools to edit the attributes of the face, such as enhancing / weakening / changing gender, getting older or younger, strengthening or weakening bags under the eyes, and big eyes Or small eyes, fatter or thinner, enhanced or weakened makeup, etc., so as to enhance business effects such as entertainment, video diversity, and live interaction. [0003] Due to the large number of attributes of the face, in order to take into account the adjustment of different attributes, the structure of the tool for editing face attributes is relatively large, and the sample size required for training is large. Moreover, running the tool with a large s...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06V40/16G06N3/045G06F18/214
Inventor 金成彬刘炉叶海佳王文斓
Owner GUANGZHOU HUYA 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