DeFMO: Deblurring and Shape Recovery of Fast Moving Objects

Objects moving at high speed appear significantly blurred when captured with cameras. The blurry appearance is especially ambiguous when the object has complex shape or texture. In such cases, classical methods, or even humans, are unable to recover the object's appearance and motion. We propose a method that, given a single image with its estimated background, outputs the object's appearance and position in a series of sub-frames as if captured by a high-speed camera (i.e. temporal super-resolution). The proposed generative model embeds an image of the blurred object into a latent space representation, disentangles the background, and renders the sharp appearance. Inspired by the image formation model, we design novel self-supervised loss function terms that boost performance and show good generalization capabilities. The proposed DeFMO method is trained on a complex synthetic dataset, yet it performs well on real-world data from several datasets. DeFMO outperforms the state of the art and generates high-quality temporal super-resolution frames.

PDF Abstract CVPR 2021 PDF CVPR 2021 Abstract

Datasets


Introduced in the Paper:

Falling Objects

Used in the Paper:

ShapeNet TbD-3D TbD
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Video Super-Resolution Falling Objects DeFMO SSIM 0.753 # 1
PSNR 26.83 # 1
TIoU 0.684 # 3
Video Super-Resolution TbD DeFMO SSIM 0.602 # 3
PSNR 25.57 # 1
TIoU 0.550 # 1
Video Super-Resolution TbD-3D DeFMO SSIM 0.699 # 1
PSNR 26.23 # 1
TIoU 0.879 # 1

Methods


No methods listed for this paper. Add relevant methods here