Super-division network model training method and device and electronic equipment

Abstract

The invention provides a super-division network model training method and device and electronic equipment. The method comprises the following steps: firstly, obtaining a plurality of sample image pairs; for each sample image pair, inputting an initial sample image in the sample image pair into the initial super-division network model to obtain a second sample image; performing at least one-scale down-sampling processing on the second sample image and the first sample image in the sample image pair to obtain a third sample image corresponding to the first sample image and a fourth sample image corresponding to the second sample image; combining the third sample image and the fourth sample image to jointly train the initial super-division network model, so that the enhancement performance of the lightweight initial super-division network model can be improved, additional network parameters are not introduced, and a lightweight target super-division network model is obtained, and the target super-division network model can better recover a high-resolution image, and the enhancement of a low-resolution image is realized, so that the reconstruction quality of the high-resolution image is improved.

Description

technical field [0001] The present application relates to the technical field of data processing, and in particular to a training method, device and electronic equipment for a super-resolution network model. Background technique [0002] Image super-resolution refers to recovering a high-resolution image from a low-resolution image. The higher the resolution of the image, the richer the image information it provides. Especially in the fields of military reconnaissance and medical diagnosis, compared with low-resolution images, high-resolution images are particularly important. [0003] The super-resolution network model can be used to restore high-resolution images corresponding to low-resolution images. However, a super-resolution network model with a larger network structure needs to learn more parameters, which can easily cause overfitting to the training data set; moreover, a network with a large number of layers requires more computing resources, and its calculation v...

AI Technical Summary

Problems solved by technology

However, the super-resolution network model with a larger network structure needs to learn more parameters, which is easy to cause overfitting to the training data set; moreover, a network with a large number of layers requires more computing resources, and its calculation volume increases quadratically

However, the super-resolution network model with a smaller network structure cannot fully learn the mapping relationship between low-resolution images and high-resolution images. Therefore, the existing super-resolution network model with a smaller network structure cannot restore low-resolution images well High-resolution images corresponding to high-resolution images

Images