An image super-resolution enhancement method based on knowledge distillation

Abstract

The invention discloses an image super-resolution enhancement method based on knowledge distillation, which includes the following steps: 1) acquisition of training data and test data; 2) training of a teacher network; the teacher network has a neural network model with a deep convolution layer , 3) The training of the student network; 4) The teacher network guides the learning of the student network; through three groups of guided experiments, the student network learns and absorbs the feature map of the teacher network; 5) Test and evaluate the image reconstruction effect; 6) According to the output feature map The different matrix relationships among them further guide the student network. The present invention uses the relevant ideas of knowledge distillation to transfer the performance of the teacher network to the student network. The student network model can efficiently run on mobile devices and embedded devices limited by low power consumption. On the premise that the network structure of the student network remains unchanged, After being guided by the teacher network, the PSNR of the student network has been significantly improved, and better reconstruction results have been obtained.

Description

technical field [0001] The invention relates to the fields of computer vision and deep learning, in particular to an image super-resolution enhancement method based on knowledge distillation. Background technique [0002] Super-resolution (Super Resolution, SR) is a classic problem in computer vision. The purpose of single image super-resolution (Single Image Super-Resolution, SISR) is to use methods such as digital image processing, from a single low-resolution (Low- The corresponding high-resolution (High-Resolution, HR) image is recovered from the Resolution, LR) image. In the super-resolution problem, assuming a low-resolution image X, our goal is to recover a super-resolution image Y′ that is as similar as possible to the ground truth (Ground Truth, GT) image Y. [0003] Traditional interpolation-based magnification methods, including Bilinear Interpolation and Bicubic Interpolation, use fixed calculation formulas to perform weighted averages using neighborhood pixel i...

AI Technical Summary

Problems solved by technology

Although increasing the depth of the network can bring better super-resolution reconstruction results, but at the same time the amount of computation and memory consumption will also increase. In many practical application scenarios (such as mobile terminals and embedded low-power consumption constraints ), unable to run deep convolutional neural network models in real time

Images