Non-reference image quality evaluation method and device under meta-learning and computer equipment

Abstract

The invention relates to a no-reference image quality evaluation method and device under meta-learning, computer equipment and a storage medium. The method comprises the steps of obtaining an input training image set; respectively extracting an edge feature map, a local texture feature map and a visual sensitivity distribution schematic diagram of the images in the image set; taking the training image set, the edge feature map, the local texture feature map, the visual sensitivity distribution schematic diagram, the brightness map and the chromaticity diagram as data sets of six quality-related tasks, and learning a quality prior model by utilizing a meta-learning framework; obtaining an input target task image set; and carrying out fine adjustment training on the quality prior model by utilizing the target task image set to obtain a final quality evaluation model. According to the method, a robust quality prior model is learned through a plurality of quality-related tasks by using a meta-learning method, and then a small number of annotation samples of a target quality evaluation task are input for fine adjustment training to obtain a final quality evaluation model, so that generalization can be quickly achieved.

Description

technical field [0001] The present invention relates to the technical field of image quality evaluation, in particular to a reference-free image quality evaluation method, device, computer equipment and storage medium under meta-learning. Background technique [0002] Image is an important carrier for people to obtain information in daily life, and various types of distortion will be introduced in the process of image acquisition, compression and transmission. The objective evaluation method of image quality can automatically evaluate the image quality while maintaining the consistency with human perception, which has important application value in the design and optimization of many image-driven related systems. [0003] At present, the objective quality evaluation can be divided into: full-reference image quality evaluation, semi-reference image quality evaluation and no-reference image quality evaluation according to the degree of dependence on reference images. The full...

AI Technical Summary

Problems solved by technology

Due to the high cost of data labeling in quality assessment databases, image quality assessment is a typical small-sample learning problem

However, the existing image quality evaluation models directly constructed using deep convolutional neural networks are prone to overfitting problems, resulting in unsatisfactory scalability of the model.

Images