Learning to Kindle the Starlight
This work addresses a domain-specific problem for astrophotography enthusiasts and researchers by providing a novel dataset and method for star field image enhancement, though it is incremental as it adapts existing diffusion models to a new application.
The paper tackles the challenge of enhancing star field images, which are difficult to capture due to light pollution and hardware requirements, by constructing the first Star Field Image Enhancement Benchmark (SFIEB) with 355 real and 854 semi-synthetic images and proposing StarDiffusion, a method based on conditional DDPM that outperforms state-of-the-art low-light image enhancement algorithms.
Capturing highly appreciated star field images is extremely challenging due to light pollution, the requirements of specialized hardware, and the high level of photographic skills needed. Deep learning-based techniques have achieved remarkable results in low-light image enhancement (LLIE) but have not been widely applied to star field image enhancement due to the lack of training data. To address this problem, we construct the first Star Field Image Enhancement Benchmark (SFIEB) that contains 355 real-shot and 854 semi-synthetic star field images, all having the corresponding reference images. Using the presented dataset, we propose the first star field image enhancement approach, namely StarDiffusion, based on conditional denoising diffusion probabilistic models (DDPM). We introduce dynamic stochastic corruptions to the inputs of conditional DDPM to improve the performance and generalization of the network on our small-scale dataset. Experiments show promising results of our method, which outperforms state-of-the-art low-light image enhancement algorithms. The dataset and codes will be open-sourced.