TY  - RPRT
TI  - Unsupervised object segmentation in video by efficient selection of highly probable positive features
AU  - Emanuela Haller
AU  - Marius Leordeanu
PY  - 2017
UR  - https://arxiv.org/abs/1704.05674
ID  - 1704.05674
ER  -