Image fixed value impulse noise denoising method and training method for image fixed value impulse noise denoising model

Abstract

The invention discloses a training method for an image fixed value impulse noise denoising model. The method comprises the following steps: estimating fixed value impulse noise density in a digital image through a noise density estimation network; obtaining a dual-channel graph based on a noise image and a noise density graph; inputting the dual-channel graph into a convolutional neural network; and learning a mapping relation with a clean image so as to train a denoising model for image fixed value impulse noise. According to the fixed value impulse noise denoising method, the trained fixed value impulse noise denoising model is combined with a noise mark matrix, noise pixel calibration is fused into the denoising network model, only noise-containing positions of the image are recovered in combination with the clean image, the integrity of noise-free information is ensured, and the restoration accuracy of noise-containing information is improved. Therefore, the network model trained by the method is superior to a traditional algorithm and an existing pulse noise reduction network in key evaluation indexes such as peak signal-to-noise ratios and structural similarity, has a good denoising effect on noise images of various concentrations, and is excellent in robust performance.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to an image fixed value impulse noise denoising method and a model training method thereof. Background technique [0002] In the process of digital image acquisition and transmission, due to the influence of equipment and environmental factors, the image is interfered by various signals, and then various noises are generated. Among them, fixed-value impulse noise is a common noise, which is reflected in black on the picture. Dots and white dots, a type of bipolar noise, appear randomly anywhere in the picture. The existence of noise makes some details of the picture lost, which not only seriously affects the visual quality of the picture, but also affects its application in the field of computer vision, such as target detection, image segmentation, medical images and remote sensing images, etc. Existing traditional filtering algorithms can cause problems such as blurred pictures and...

AI Technical Summary

Problems solved by technology

The denoising ability of existing deep learning algorithms is better than that of traditional algorithms, but a well-trained model can only be effective for specific concentrations of noise, and its generalization ability is poor

Images