In this paper, we present two content-adaptive rain and snow removal algorithms for single image based on filtering. The first algorithm treats rain and snow removal task as an issue of bilateral filtering, where a content-based saliency prior is introduced. While the other views the same task from the perspective of guided-image-filtering, and the guidance image is derived according to the statistical property of raindrops or snowflakes as well as image background content. A comparative study and quantitative evaluation with some main existing image assessment algorithms demonstrate better performance of our proposed algorithms. The main contributions of our works are twofold: firstly, to the best of our knowledge, our algorithms are among the first to introduce image content information for single-image-based rain and snow removal; and secondly, we are also among the first to introduce quantitative assessment for single-image-based rain and snow removal tasks.
CITATION STYLE
Yu, S., Zhao, Y., Mou, Y., Wu, J., Han, L., Yang, X., & Zhao, B. (2014). Content-adaptive rain and snow removal algorithms for single image. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8866, pp. 439–448). Springer Verlag. https://doi.org/10.1007/978-3-319-12436-0_49
Mendeley helps you to discover research relevant for your work.