Model training method and media information synthesis method and device

A technology for training information and model training, applied in character and pattern recognition, image data processing, instruments, etc., can solve problems such as image content is not smooth, reducing the authenticity of synthetic videos, etc.

Active Publication Date: 2020-02-28
TENCENT TECH (SHENZHEN) CO LTD
View PDF3 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, due to the difference between the region generated by using the general hint and the real background, the image content will not be smooth when splicing, so that the synthesized video will have more obvious splicing defects, reducing the quality of the synthesized video. authenticity

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
  • Model training method and media information synthesis method and device
  • Model training method and media information synthesis method and device
  • Model training method and media information synthesis method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0120] The embodiment of the present application provides a method for model training, a method and a device for synthesizing media information, which can use a content mask to generate an image to be trained, and the coverage of the content mask is larger than the range of the object to be trained, so that the boundary is properly defined. The expansion can make the part between the image content and the background generated by the information synthesis model, avoiding the defects caused by splicing, thereby improving the authenticity of the synthesized video.

[0121] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above drawings are used to distinguish similar objects, and not necessarily Used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the application described herein,...

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 model training method and a media information synthesis method and device, a to-be-trained image can be generated by using a content mask, and the coverage range of the content mask is greater than the range of a to-be-trained object, so that a part between the image content and the background can be generated by an information synthesis model, and the authenticity of a synthesized video is improved. The method provided by the invention comprises the following steps: acquiring an image sample set and general prompt information, obtaining an image sample set accordingto the image sample set and general prompt information; and generating a content mask set, generating a to-be-trained image set according to the content mask set, obtaining a prediction image set through the to-be-trained information synthesis model based on the image sample set and the to-be-trained image set, and training the to-be-trained information synthesis model by adopting a target loss function to obtain an information synthesis model.

Description

technical field [0001] The present application relates to the field of artificial intelligence, in particular to a method for model training, a method and a device for synthesizing media information. Background technique [0002] With the diversified development of media forms, more and more users start to use mobile terminals to take videos, pictures or record audio to record every detail of work and life. For the scene of video synthesis, the user can integrate the face or human body of the target person into a specified background, and when the video is played, the target person can be seen appearing in the specified background. [0003] At present, the general prompt information of the face of the target person can be used, or the general prompt information of the human body part can be used to directly generate the region with the general prompt. The rough hint usually comes from a rendering result of a face or a human body, a texture map, or coordinates of key points,...

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): G06T3/00G06K9/62
CPCG06T3/0012G06F18/214
Inventor 黄浩智李佳维沈力凌永根刘威俞栋
Owner TENCENT TECH (SHENZHEN) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap