{
  "id": "1711.07064",
  "version": "v1",
  "published": "2017-11-19T19:46:18.000Z",
  "updated": "2017-11-19T19:46:18.000Z",
  "title": "DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networks",
  "authors": [
    "Orest Kupyn",
    "Volodymyr Budzan",
    "Mykola Mykhailych",
    "Dmytro Mishkin",
    "Jiri Matas"
  ],
  "categories": [
    "cs.CV"
  ],
  "abstract": "We present an end-to-end learning approach for motion deblurring, which is based on conditional GAN and content loss. It improves the state-of-the art in terms of peak signal-to-noise ratio, structural similarity measure and by visual appearance. The quality of the deblurring model is also evaluated in a novel way on a real-world problem -- object detection on (de-)blurred images. The method is 5 times faster than the closest competitor. Second, we present a novel method of generating synthetic motion blurred images from the sharp ones, which allows realistic dataset augmentation. Model, training code and dataset are available at https://github.com/KupynOrest/DeblurGAN",
  "revisions": [
    {
      "version": "v1",
      "updated": "2017-11-19T19:46:18.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "conditional adversarial networks",
      "blind motion deblurring",
      "generating synthetic motion blurred images",
      "structural similarity measure",
      "realistic dataset augmentation"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}