Publications

ImageCBMI Jordi Pont-Tuset, Miquel A. Farré, Aljoscha Smolic
Semi-Automatic Video Object Segmentation by Advanced Manipulation of Segmentation Hierarchies
International Workshop on Content-Based Multimedia Indexing (CBMI), June 2015
[PDF] [BibTeX]
@inproceedings{PFS2015,
author = {Pont-Tuset, J. and Farr\'{e}, M.A. and Smolic, A.},
title = {Semi-Automatic Video Object Segmentation by Advanced Manipulation of Segmentation Hierarchies},
booktitle = {International Workshop on Content-Based Multimedia Indexing (CBMI)},
year = {2015}
}

Dataset


[23-Aug-2015] We did a minot update in the dataset to remove some empty annotations.
You can download the original videos from the BVSD [1] site (train and test) or from the Video Segmentation Benchmark (VSB100) [2] site.

We provide object annotations for 100 high-definition video sequences from the Berkeley Video Segmentation Dataset (BVSD) [3]. We analized the segmentation annotations provided in the Video Segmentation Benchmark (VSB100) [2] to find regions in which the annotations from the diverse individuals were coherent and extracted the object-level annotations. Having various annotations per frame on the same object allows us to quantify the difficulty of each object by comparing them all-vs-all. We get a mean value of Jaccard (Intersection over Union) of J = 0.89 ± 0.08. Here some example of discrepancies between annotations and it's Jaccard values:
AnnotationExample

If you use this database please cite:
[1] Jordi Pont-Tuset, Miquel A. Farré, Aljoscha Smolic;
Semi-Automatic Video Object Segmentation by Advanced Manipulation of Segmentation Hierarchies,
International Workshop on Content-Based Multimedia Indexing (CBMI), 2015.
[2] F. Galasso, N.S. Nagaraja, T.J. Cardenas, T. Brox, B. Schiele;
A Unified Video Segmentation Benchmark: Annotation, Metrics and Analysis,
International Conference on Computer Vision (ICCV), 2013.
[3] P. Sundberg, T. Brox, M. Maire, P. Arbelaez, and J. Malik;
Occlusion Boundary Detection and Figure/Ground Assignment from Optical Flow,
Computer Vision and Pattern Recognition (CVPR), 2011.