Human face restoration model training method, restoration method, device and equipment and medium

A technology for repairing models and training methods, which is applied in the field of artificial intelligence-based face repairing, which can solve problems such as difficult construction and acquisition, and unsatisfactory repairing effects, so as to achieve normalization, improve repairing effects, and reduce errors.

Pending Publication Date: 2020-08-07
BEIJING BAIDU NETCOM SCI & TECH CO LTD
View PDF16 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, the problem existing in the existing technology is that it is difficult to construct and obtain real low-quality face image and high-definition face image sample pairs. to construct a sample pair
Therefore, th

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
  • Human face restoration model training method, restoration method, device and equipment and medium
  • Human face restoration model training method, restoration method, device and equipment and medium
  • Human face restoration model training method, restoration method, device and equipment and medium

Examples

Experimental program
Comparison scheme
Effect test

Example

[0093] First embodiment

[0094] figure 1 It is a schematic flow chart of a method for training a face restoration model according to the first embodiment of the present application. This embodiment is used in the case of model training for restoration of a first-quality face image. The method can be The training device of the face repair model can be implemented. The device can be implemented in software and / or hardware, and can be integrated into electronic equipment with specific computing capabilities. Such as figure 1 As shown, the training method of a face repair model provided in this embodiment may include:

[0095] S110. Acquire a sample pair of a first-quality face image and a second-quality face image, where the second-quality face image of the sample pair is used as a supervision image.

[0096] Wherein, the face image of the first image quality and the face image of the second image quality are face images of different image quality, and the perceived quality of the fir...

Example

[0120] Second embodiment

[0121] Figure 4 It is a schematic flow chart of a method for training a face repair model according to the second embodiment of the present application. This embodiment is a further optimization of the above-mentioned embodiment. Such as Figure 4 As shown, the training method of a face repair model provided in this embodiment may include:

[0122] S410. Acquire a sample pair of a first-quality face image and a second-quality face image, where the second-quality face image of the sample pair is used as a supervision image.

[0123] S420: Input the first quality face image of the sample pair into the decompression model to perform compression noise removal processing.

[0124] Among them, after the low-quality face image is obtained, the low-quality face image is input to the decompression model, and the low-quality face image is subjected to compression noise removal processing. The decompression model can remove the blocks in the low-quality face image Sh...

Example

[0134] The third embodiment

[0135] Image 6 It is a schematic flow chart of a method for repairing a face image according to the third embodiment of the present application. This embodiment is used in the case of generating a second-quality face image based on a first-quality face image. The method can be A facial image restoration device is implemented, which can be implemented in software and / or hardware, and can be integrated into electronic equipment with computing capabilities. Such as Image 6 As shown, a method for repairing a face image provided in this embodiment may include:

[0136] S610: Acquire a face image of the first quality to be repaired.

[0137] Wherein, the first-quality face image to be repaired may be a low-quality image to be repaired, and the low-quality image to be repaired is an input image.

[0138] S620. Input the first-quality face image to be repaired into a face repair model for processing, where the face repair model is obtained by training using th...

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 discloses a training method and device of a face restoration model, a restoration method and device of the face restoration model, equipment and a medium, and relates to the technical field of image processing, in particular to the face restoration technology based on artificial intelligence. According to the specific implementation scheme, a sample pair of a firstimage quality face image and a second image quality face image is obtained, and the second image quality face image of the sample pair serves as a supervision image; inputting the first image qualityface image of the sample pair into a face restoration model for training; based on at least two loss functions, respectively calculating at least two loss relationships between an output image and a supervision image of the face restoration model; if it is determined that at least two loss relations do not meet the set convergence requirement, model parameters of the face restoration model are adjusted, training is continued until it is determined that at least two loss relations meet the set convergence requirement, and it is determined that training of the face restoration model is completed. And the face restoration model is trained through the sample pair and at least two loss functions, so that the accuracy of face restoration is improved.

Description

technical field [0001] The embodiments of the present application relate to the technical field of image processing, and in particular to artificial intelligence-based face restoration technology. Background technique [0002] With the rapid development of image restoration technology, many low-quality images can be repaired into high-quality images. An important application scenario of image restoration technology is the restoration of face images, which is used to process and restore low-quality face images to obtain clear face images. [0003] When using a machine learning model to restore face images, the main method is to use low-quality face images and high-definition face images as sample pairs, and train the machine learning model through a large number of sample pairs, so as to use the trained machine learning model Restoring face images. [0004] However, the problem existing in the existing technology is that it is difficult to construct and obtain real low-qual...

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): G06T5/00G06K9/00G06N3/04G06N3/08
CPCG06T5/003G06N3/084G06V40/168G06N3/045Y02T10/40
Inventor 丁予康何栋梁李超张赫男孙昊文石磊丁二锐章宏武
Owner BEIJING BAIDU NETCOM SCI & TECH 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