- python 3.6
- pytorch 1.0.1.post2
- numpy, opencv, pillow
wget -O STM_weights.pth "https://www.dropbox.com/s/mtfxdr93xc3q55i/STM_weights.pth?dl=1"
python eval_DAVIS.py -g '1' -s val -y 16 -D [path/to/DAVIS]
python eval_DAVIS.py -g '1' -s val -y 17 -D [path/to/DAVIS]
If you are not able to run our code but interested in our results. The pre-computed results will be helpful.
If you find our paper and repo useful, please cite our paper. Thanks!
Video Object Segmentation using Space-Time Memory Networks
Seoung Wug Oh, Joon-Young Lee, Ning Xu, Seon Joo Kim
ICCV 2019
Fast Video Object Segmentation by Reference-Guided Mask Propagation
Seoung Wug Oh, Joon-Young Lee, Kalyan Sunkavalli, Seon Joo Kim
CVPR 2018
A modified STM model is used for DAVIS Interactive VOS Challenge 2019 (https://davischallenge.org/challenge2019/interactive.html). If you are intersted in comparison with our interactive STM model, please use evaluation summary obtained from the DAVIS framework. [Download link (DAVIS-17-val)]. The timing is measured using a single 2080Ti GPU. We will make a demo software available soon. For the further questions, please contact me by E-mail.
This software is for non-commercial use only. The source code is released under the Attribution-NonCommercial-ShareAlike (CC BY-NC-SA) Licence (see this for details)