Temporally consistent dense video annotations are scarce and hard to collect. In contrast, image segmentation datasets (and pre-trained models) are ubiquitous, and easier to label for any novel task. In this paper, we introduce a method to adapt still image segmentation models to video in an unsupervised manner, by using an optical flow-based consistency measure. To ensure that the inferred segmented videos appear more stable in practice, we verify that the consistency measure is well correlated with human judgement via a user study. Training a new multi-input multi-output decoder using this measure as a loss, together with a technique for refining current image segmentation datasets and a temporal weighted-guided filter, we observe stability improvements in the generated segmented videos with minimal loss of accuracy.
@article{arxiv.2110.08893,
title = {Temporally stable video segmentation without video annotations},
author = {Aharon Azulay and Tavi Halperin and Orestis Vantzos and Nadav Borenstein and Ofir Bibi},
journal= {arXiv preprint arXiv:2110.08893},
year = {2022}
}