Image deblurring algorithm based on knowledge distillation and deep neural network

Abstract

The invention discloses an image deblurring algorithm based on knowledge distillation and a deep neural network, which can be used for realizing high-quality conversion from a blurred image to a clear image. The algorithm comprises the following steps: constructing a training set, a verification set and a test set according to a public image data set; constructing a self-coding deep neural network based on a Unet structure, namely a teacher model; performing supervised training on the teacher model based on the high-definition image data set until convergence; constructing a deep neural network model, namely a student model, based on a Unet structure and deformable convolution; on the basis of the complete blurred image data set, combined with a converged teacher network, supervising and training the deblurring student model by using a knowledge distillation method, and further obtaining a converged deep deblurring model; and evaluating the deblurring performance of the student network model on the blurred image test set. According to the method, better deblurring performance is obtained, and better structural similarity is obtained while a higher peak signal-to-noise ratio is obtained.

Description

technical field [0001] The invention belongs to the technical field of image deblurring in computer vision, in particular to an image deblurring algorithm based on knowledge distillation and deep neural network. Background technique [0002] Digital image is an important information carrier, which is widely used in various fields, such as: digital audio and video, remote sensing detection and traffic management. However, digital images are easily affected by many destructive factors in the process of acquiring, disseminating and storing, which in turn causes the degradation of digital image quality, and blurring is one of the main features. Therefore, the research on deblurring of digital images is of great significance and of great value. [0003] The traditional image deblurring algorithm mainly models the image blurring process as a linear convolution model. MLE) to solve a minimized energy function. However, traditional deblurring algorithms are limited to limited sce...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide an image deblurring algorithm based on knowledge distillation and deep neural network to solve the problem that the existing image deblurring algorithm does not have higher practical application value and the existing image deblurring algorithm based on deep neural network Technical issue with limited performance improvement in deblurring

Images