Image super-resolution reconstruction method based on deep learning iteration up-down sampling

Abstract

The invention relates to an image super-resolution reconstruction method based on deep learning iteration up-down sampling. The method comprises the following steps: preparing an original high resolution image, performing bicubic interpolation on the original high-resolution image to obtain a low-resolution image, inputting the low-resolution image into a constructed neural network; extracting low-resolution image features; extracting high-resolution image features layer by layer through upper-lower sampling modules of a plurality of back projection layers, so that the interdependence relationship between the low-resolution image and the high-resolution image can be mined more, wherein the convolution in the neural network adopts cavity convolution to increase a receptive field, so the sensitivity of the network to feature information is improved, dense connection is introduced to reduce information loss caused by intermediate layer transmission, low-layer feature information can be better utilized, and the performance of image reconstruction is improved.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image super-resolution reconstruction method based on deep learning iterative up-down sampling. Background technique [0002] The application field of image super-resolution reconstruction is extremely broad, and it has important application prospects in military, medical, public security, and computer vision. For example, high-resolution medical images are very helpful for doctors to make correct diagnoses; using high-resolution satellite images, it is easy to distinguish similar objects from similar objects; if high-resolution images can be provided, computer vision The performance of pattern recognition will be greatly improved. The most direct way to improve image resolution is to improve the optical hardware in the acquisition system, but this method is limited by the constraints of difficult to greatly improve the manufacturing process and high manufacturing costs. Ther...

AI Technical Summary

Problems solved by technology

Although compared with the previous traditional method, the traditional image super-resolution algorithm based on shallow learning has been greatly improved, such as sharper, but because the features extracted by this method are artificially designed, it cannot express high-level features, so super-resolution tasks with large magnifications are still difficult to apply to this method

Images