Model training method, denoising method, model, equipment and storage medium

Abstract

The invention provides a model training method, a denoising method, a model, equipment and a storage medium. In the embodiment of the invention, a first adversarial network comprising a background generator and a noise discriminator can be constructed in a denoising model, on the basis of the first adversarial network, noise features in the noise-containing image can be determined according to a background image extracted from the noise-containing image by the background generator, and the noise features are discriminated by using the noise discriminator. Therefore, through mutual confrontation between the background generator and the noise discriminator, the first constraint parameter corresponding to the background generator can be calculated, so that unsupervised training can be performed on the background generator based on the first constraint parameter. In the embodiment of the invention, the training constraint on the background generator can be determined based on the noise-containing image, so that unsupervised training of the background generator in the denoising model can be realized, and the generalization ability and precision of the background generator can be effectively improved due to the fact that the noise-containing image in the natural scene can be used as the training sample.

Description

technical field [0001] The present application relates to the technical field of machine learning, and in particular to a model training method, a denoising method, a model, a device, and a storage medium. Background technique [0002] Image De-raining technology, as an image enhancement technology in the direction of deep learning and computer vision, has shown great practical value in surveillance video and other scenarios. [0003] At present, since it is difficult to collect clean and rainy training samples that appear in pairs in real scenes, it is necessary to use techniques such as PS synthesis to add raindrops to the original clean images, and then rely on rain-clean paired images as training set for fully supervised training of the model. However, the characteristics of synthetic raindrops are quite different from those in real scenes, which leads to insufficient generalization ability of the model produced by this training method, and the deraining accuracy in rea...

AI Technical Summary

Problems solved by technology

However, the characteristics of synthetic raindrops and raindrops in real scenes are quite different, resulting in insufficient generalization ability of the model produced by this training method, and the rain removal accuracy in real natural scenes is too low

Images