Gyroscope-Aided Motion Deblurring with Deep Networks

Janne Mustaniemi, Juho Kannala, Simo Särkkä, Jiri Matas, Janne Heikkilä

WACV 2019 pp. 1914-1922

doi:10.1109/WACV.2019.00208 /wacv/2019/mustaniemi2019wacv-gyroscope/

Abstract

We propose a deblurring method that incorporates gyroscope measurements into a convolutional neural network (CNN). With the help of such measurements, it can handle extremely strong and spatially-variant motion blur. At the same time, the image data is used to overcome the limitations of gyro-based blur estimation. To train our network, we also introduce a novel way of generating realistic training data using the gyroscope. The evaluation shows a clear improvement in visual quality over the state-of-the-art while achieving real-time performance. Furthermore, the method is shown to improve the performance of existing feature detectors and descriptors against the motion blur.

PDF WACV Semantic Scholar

Cite

Text

Mustaniemi et al. "Gyroscope-Aided Motion Deblurring with Deep Networks." IEEE/CVF Winter Conference on Applications of Computer Vision, 2019. doi:10.1109/WACV.2019.00208

Markdown

[Mustaniemi et al. "Gyroscope-Aided Motion Deblurring with Deep Networks." IEEE/CVF Winter Conference on Applications of Computer Vision, 2019.](https://mlanthology.org/wacv/2019/mustaniemi2019wacv-gyroscope/) doi:10.1109/WACV.2019.00208

BibTeX

@inproceedings{mustaniemi2019wacv-gyroscope,
  title     = {{Gyroscope-Aided Motion Deblurring with Deep Networks}},
  author    = {Mustaniemi, Janne and Kannala, Juho and Särkkä, Simo and Matas, Jiri and Heikkilä, Janne},
  booktitle = {IEEE/CVF Winter Conference on Applications of Computer Vision},
  year      = {2019},
  pages     = {1914-1922},
  doi       = {10.1109/WACV.2019.00208},
  url       = {https://mlanthology.org/wacv/2019/mustaniemi2019wacv-gyroscope/}
}