Face image super-resolution calculation method based on deep learning

Abstract

The invention relates to the technical field of image recognition, in particular to a face image super-resolution calculation method based on deep learning. The method comprises the steps of 1, data processing; a reference image-based face super-resolution method is used for inputting data as a low-resolution image LR and a high-resolution reference image Ref, and an output target is a super-resolution image SR; 2, texture information is extracted; 3, texture information conversion; the texture information conversion module comprises a correlation coefficient matrix calculation module, a texture information query module and a texture weight assignment module; step 4, super-resolution processing based on the reference image; step 5, a loss function and an optimizer; step 6, evaluation method; step 7, verifying the effectiveness of the TFSR; and comparing the TFSR with the optimal face super-resolution method to verify the effectiveness of the TFSR.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a deep learning-based face image super-resolution calculation method. Background technique [0002] In the existing deep learning-based FSR algorithms, it is a very popular and effective practice to integrate face prior knowledge (such as face key points, analytical maps, and identity information) into face super-resolution. The existing technical solution is DIC, which solves the following two problems: [0003] 1. The quality of LR face images or preliminary SR reconstruction results is preferred, and the accuracy of the relevant information extracted from them is difficult to guarantee; [0004] 2. Most of the methods just use the feature map in parallel to utilize the prior information of the face, which cannot make full use of the prior information. [0005] To address the problem of inaccurate face priors from low-quality images, DIC uses figure 1 The shown fee...

AI Technical Summary

Problems solved by technology

[0003] 1. The quality of LR face images or preliminary SR reconstruction results is prioritized, and the accuracy of relevant information extracted from them is difficult to guarantee;

[0004] 2. Most of the methods only use the feature map to use the prior information of the face, which cannot make full use of the prior information

[0007] DIC recovers the prior information of the face from the super-resolution result, but the super-resolution result is not completely consistent with the texture details of the high-resolution target, so the prior information is inaccurate

Correspondingly, the accuracy of super-resolution results derived from inaccurate prior information cannot be guaranteed

Images