Low-illumination image enhancement method based on attention mechanism and multi-level feature fusion

Abstract

The invention relates to a low-illumination image enhancement method based on an attention mechanism and multi-level feature fusion, and the method comprises the following steps: carrying out the processing of a low-illumination image at an input end, and outputting a four-channel feature map; using a convolutional layer based on an attention mechanism as a feature extraction module and for extracting basic features as low-level features; fusing the low-level features with the corresponding high-level features and the deepest feature of the convolutional layer, and obtaining a final feature map after deconvolution; and outputting the mapping to restore the final feature map into the RGB picture. According to the invention, multi-level features of the deep convolutional neural network model are fully utilized. Different levels of features are fused, different weights are given to a feature channel through a channel attention mechanism, better feature representation is obtained, the image processing accuracy is improved, a high-quality image is obtained, and the method can be widely applied to the technical field of computer low-level visual tasks.

Description

technical field [0001] The invention belongs to the technical field of computer image processing, in particular to a low-illuminance image enhancement method based on attention mechanism and multi-level feature fusion. Background technique [0002] In the field of computer image processing, low-illuminance image enhancement technology refers to processing low-contrast and illuminated images by a certain method, and finally obtaining a clear image with high brightness. The processed output images can be widely used in high-level vision tasks, such as object detection, pedestrian re-identification, and automatic driving. [0003] Traditional low-light image enhancement methods mainly use histogram equalization and Retinex-based methods. In recent years, with the development of machine learning and big data, image enhancement algorithms based on deep neural networks have made great progress. Low-light image enhancement has gradually become one of the research hotspots in comp...

AI Technical Summary

Problems solved by technology

[0004] Although low-light image enhancement methods have been greatly developed, there are still many problems to be solved due to the difficulty of low-light images.

The difficulties of low-light image enhancement are mainly reflected in: (1) The uncertainty of the low-light image itself. Due to the low light intensity of the shooting environment and some non-subjective factors such as shooting shake, the low-light image contains more (2) The number of data sets is small, and it is difficult to shoot low-light images and their corresponding comparison images in real life. Therefore, the processing method based on deep convolutional neural network has a small amount of data, and it is difficult to obtain a better one. The training effect; (3) The choice of network structure

[0005] To sum up, the current method based on deep convolutional neural network is still being explored, and no suitable method has been found to process low-light images.

Images