Reblur2Deblur: Deblurring Videos via Self-Supervised Learning
This addresses motion blur in videos for computer vision applications, offering an incremental enhancement over existing methods.
The paper tackles video deblurring by fine-tuning existing neural networks with a self-supervised approach that enforces consistency between output and input blur, resulting in significant improvements in visual quality and metrics on multiple datasets.
Motion blur is a fundamental problem in computer vision as it impacts image quality and hinders inference. Traditional deblurring algorithms leverage the physics of the image formation model and use hand-crafted priors: they usually produce results that better reflect the underlying scene, but present artifacts. Recent learning-based methods implicitly extract the distribution of natural images directly from the data and use it to synthesize plausible images. Their results are impressive, but they are not always faithful to the content of the latent image. We present an approach that bridges the two. Our method fine-tunes existing deblurring neural networks in a self-supervised fashion by enforcing that the output, when blurred based on the optical flow between subsequent frames, matches the input blurry image. We show that our method significantly improves the performance of existing methods on several datasets both visually and in terms of image quality metrics. The supplementary material is https://goo.gl/nYPjEQ