CPC G06T 5/90 (2024.01) [G06T 5/20 (2013.01); G06V 10/761 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/20208 (2013.01)] | 19 Claims |
1. An image processing method, wherein the method comprises:
acquiring an original image;
performing a fuzzy processing to the original image to obtain a fuzzy image;
generating a first characteristic matrix by performing a high-dynamic-range processing to the original image using a first network model obtained by pre-training, wherein the first network model comprises a dense residual module and a gate-control-channel conversion module, the dense residual module comprises a dense-connection network and a residual network, an input of the residual network comprises output results of a plurality of convolutional layers in the dense-connection network; the gate-control-channel conversion module is configured to analyze differences between a plurality of input characteristic channels, and determine a weight of each of the characteristic channels according to analysis results;
obtaining an auxiliary characteristic matrix of the original image according to the fuzzy image, wherein the auxiliary characteristic matrix comprises at least one of the following information: detail information of the original image and low-frequency information of the original image; and
obtaining a target image according to the first characteristic matrix and the auxiliary characteristic matrix.
|