Realistic Blur Synthesis for Learning Image Deblurring
This addresses the challenge of limited and unrealistic training data for image deblurring models, which is incremental as it builds on existing synthetic and real datasets.
The paper tackled the problem of unrealistic synthetic blur datasets for training deblurring models by analyzing factors causing differences between real and synthetic blur, presenting the RSBlur dataset for analysis, and developing a novel blur synthesis pipeline that improves deblurring performance on real images.
Training learning-based deblurring methods demands a tremendous amount of blurred and sharp image pairs. Unfortunately, existing synthetic datasets are not realistic enough, and deblurring models trained on them cannot handle real blurred images effectively. While real datasets have recently been proposed, they provide limited diversity of scenes and camera settings, and capturing real datasets for diverse settings is still challenging. To resolve this, this paper analyzes various factors that introduce differences between real and synthetic blurred images. To this end, we present RSBlur, a novel dataset with real blurred images and the corresponding sharp image sequences to enable a detailed analysis of the difference between real and synthetic blur. With the dataset, we reveal the effects of different factors in the blur generation process. Based on the analysis, we also present a novel blur synthesis pipeline to synthesize more realistic blur. We show that our synthesis pipeline can improve the deblurring performance on real blurred images.