Not Just Streaks: Towards Ground Truth for Single Image Deraining
This addresses the sim2real domain gap in image deraining for computer vision applications, though it is incremental as it builds on existing deraining methods.
The authors tackled the lack of real-world paired datasets for single image deraining by collecting a large-scale dataset of real rainy and clean image pairs, enabling quantitative evaluation and training. Their proposed deep neural network outperformed state-of-the-art methods on real rainy images under various conditions.
We propose a large-scale dataset of real-world rainy and clean image pairs and a method to remove degradations, induced by rain streaks and rain accumulation, from the image. As there exists no real-world dataset for deraining, current state-of-the-art methods rely on synthetic data and thus are limited by the sim2real domain gap; moreover, rigorous evaluation remains a challenge due to the absence of a real paired dataset. We fill this gap by collecting a real paired deraining dataset through meticulous control of non-rain variations. Our dataset enables paired training and quantitative evaluation for diverse real-world rain phenomena (e.g. rain streaks and rain accumulation). To learn a representation robust to rain phenomena, we propose a deep neural network that reconstructs the underlying scene by minimizing a rain-robust loss between rainy and clean images. Extensive experiments demonstrate that our model outperforms the state-of-the-art deraining methods on real rainy images under various conditions. Project website: https://visual.ee.ucla.edu/gt_rain.htm/.