We introduce the largest real-world image deblurring dataset constructed from smartphone slow-motion videos.
By capturing 240 frames over one second, we simulate realistic long-exposure blur by averaging frames to produce blurry images, while using the temporally centered frame as the sharp reference.
Our dataset contains over 42,000 high-resolution blur-sharp image pairs, making it approximately 10 times larger than widely used datasets, with 8 times the amount of different scenes, including indoor and outdoor environments, with varying object and camera motions.
We benchmark multiple state-of-the-art (SOTA) deblurring models on our dataset and observe significant performance degradation, highlighting the complexity and diversity of our benchmark.
Our dataset serves as a challenging new benchmark to facilitate robust and generalizable deblurring models.
Blogger's Review: The release of this dataset marks a significant advancement in the field of deblurring, especially for real-world applications.
With a rich variety of image pairs, researchers can better train models to enhance the practicality and reliability of deblurring technologies.
The establishment of such datasets provides a solid foundation for future research, aiding the progress of image processing techniques.