Bayesian Semantic Instance Segmentation in Open Set World

Demo

In this demo, we use images from coco mini val dataset. We have already pre-computed object detections, ucms and (hierarchical) region features

• Run setup.m
• Run demo.m

Test on new images.

• Create a new folder under datasets, and save images under datasets/abc/images (Follow the structure of ms_coco_samples for reference.).
• Run COB network https://github.com/kmaninis/COB on these images to get ucms, and save them under ucms folder.
• Run an object detector network to generate bounding box detections, and save them under detections folder.
• Run compute_tree_and_feautures.m to compute region hierarchies and features.
• Update demo.m script with the new path to your dataset, and then run demo.m.

Citation:

If you use this code, please consider citing the following papers:

@Inproceedings{Pham2018,
   author = {{Pham}, et. al.},
    title = "{Bayesian Semantic Instance Segmentation in Open Set World}",
    booktitle={ECCV}, 
    year={2018}, 
}

If you encounter any problems with the code, please contact the first author at trungptt at gmail dot com. Enjoy!