Image blind motion deblurring method based on CNN-Transform hybrid auto-encoder

Abstract

An image blind motion deblurring method based on a CNN-Transform hybrid auto-encoder comprises two stages, namely a model training stage and a prediction stage, and the model training stage comprises the following steps: 1, preparing an image deblurring standard data set; 2, preprocessing experimental data; 3, inputting a blurred picture in a training set of the image deblurring standard data set into a hybrid auto-encoder part for recovery; 4, inputting the potential clear image generated by the hybrid auto-encoder and the corresponding target clear image in the standard data set into a discriminator, and returning the gradient after calculating the loss by the discriminator; and 5, receiving the gradient from the discriminator by the hybrid auto-encoder to perform parameter updating, wherein the model prediction stage comprises a step of inputting a blurred picture into a trained hybrid auto-encoder, and outputting the blurred picture as a deblurred clear picture. By using the method, a better image deblurring effect can be obtained, and image details can be recovered more clearly.

Description

technical field [0001] The invention belongs to the technical field of computer vision and image processing, and relates to an image blind motion deblurring method based on a CNN-Transformer hybrid autoencoder. Background technique [0002] As an important medium for transmitting information, images play an indispensable role in people's life and work. But the problem of image blurring is ubiquitous in life: due to problems such as camera shake, fast movement of objects or out of focus, the image quality will be reduced and the image will become blurred. Furthermore, various reasons: such as depth variation, occlusion of motion boundaries complicate blurring. Many shooting scenes are constantly changing over time and are not repeatable. Camera shake or blur caused by fast moving objects will most likely result in unusable images. How to effectively use these blurred images, the research on image deblurring methods is of great significance. [0003] Image deblurring has be...

AI Technical Summary

Problems solved by technology

However, most methods do not perform well enough in the balance between spatial texture details and high-level contextual information, and the visual effect is not good enough.

Images