Fork me on GitHub

An extended version of this work can be found here

OSVOS: One-Shot Video Object Segmentation

State-of-the-Art Results in Accuracy and Speed

Publication

S. Caelles*, K.K. Maninis*, J. Pont-Tuset, L. Leal-Taixé, D. Cremers, and L. Van Gool
One-Shot Video Object Segmentation, Computer Vision and Pattern Recognition (CVPR), 2017.
[BibTeX] [PDF]
@inproceedings{Cae+17,
  Author 	= {Sergi Caelles and Kevis-Kokitsi Maninis and Jordi Pont-Tuset and Laura Leal-Taix\'e and Daniel Cremers and Luc {Van Gool}},
  Title 	= {One-Shot Video Object Segmentation},
  Booktitle	= {Computer Vision and Pattern Recognition (CVPR)},
  Year 		= {2017}
}

Results

The table shows the overall results of OSVOS compared to the state of the art in the validation set of DAVIS.
We are 11.8 points above OFL.
Below the per-sequence results of OSVOS compared to the previous state of the art.
In terms of speed, below the plot of quality versus time per frame.
We reach the same quality than OFL in 160 miliseconds vs 40 seconds.
At the same speed than BVS, we are 15 points better.

For a complete comparison and qualitative results, visit the DAVIS website.

Downloads