Image processing method and image processing system

Abstract

The invention provides an image processing method and an image processing system. The image processing method includes the following steps: receiving an original image, and reducing the size of the original image to obtain a first transitional image; providing a first expansion convolutional neural network, and using the first expansional convolutional neural network to perform expansion and convolution on the first transitional image Product processing, to enhance the quality of the first transition image, to obtain the second transition image; enlarge the size of the second transition image, to obtain the third transition image, the size of the third transition image is the same as the size of the original image; the original image and The third transitional image matrix is ​​spliced ​​to obtain the fourth transitional image; a second expansion convolutional neural network is provided, and the second expansional convolutional neural network is used to perform expansion convolution processing on the fourth transitional image to restore the details of the fourth transitional image , to get the target image. The invention can reduce the amount of calculation of the neural network without reducing the detail information of the image, and avoid local mutations in the processed image.

Description

technical field [0001] The invention relates to the field of display technology, in particular to an image processing method and an image processing system. Background technique [0002] With the development of science and technology, people have higher and higher requirements for display devices, and lighter, clearer and more colorful TVs have become the development direction and goal of TVs. In addition to improving the TV resolution on the hardware, increasing the dynamic range and color gamut range, etc., the image signal can be optimized one by one through the image processing system set in the display device to direct and coordinate the functions of the display device, so that it can be used in the existing On the basis of hardware, it brings better picture quality to users, [0003] Image enhancement technology is a kind of image processing technology, which can significantly improve image quality, make image content more layered and subjective observation effect mor...

AI Technical Summary

Problems solved by technology

[0006] Due to the characteristics of the expanded convolution, the expanded convolutional neural network can maintain a small number of calculations while expanding the field of view of the neural network, and realize the output of the neural network with the pixel accuracy of the original image. However, for large-scale images, the expanded convolutional neural network There is a problem of local mutation of the image due to insufficient field of view. For this problem, the existing technology is forced to deepen the network depth, resulting in redundant network capacity and increased complexity of the network architecture.

Images