Perception Evaluation: A New Solar Image Quality Metric Based on the Multi-fractal Property of Texture Features

The next generation of ground-based solar observations requires good image quality metrics for post facto processing techniques. Based on the assumption that texture features in solar images are multi-fractals and that they can be extracted by trained deep neural networks as feature maps, a new reduced-reference objective image quality metric, the perception evaluation, is proposed. The perception evaluation is defined as the cosine distance of the Gram matrix between feature maps extracted from high-resolution reference images and that from blurred images. We evaluate the performance of the perception evaluation using simulated blurred images and real observed images. The results show that with a high-resolution image as reference, the perception evaluation can give a robust estimate of the image quality for solar images in different scenarios.