Object Segmentation Tracking from Generic Video Cues

TL;DR

This paper introduces a lightweight variational method for online video object segmentation using optical flow and image boundaries, offering a competitive alternative to CNN-based approaches with lower computational cost.

Contribution

The authors present a novel variational framework that leverages generic video cues for object segmentation, reducing reliance on expensive CNN training.

Findings

Competitive segmentation accuracy on DAVIS and SegTrack datasets.

Can enhance CNN-based segmentation results when combined.

Operates efficiently with minimal parameter tuning.

Abstract

We propose a light-weight variational framework for online tracking of object segmentations in videos based on optical flow and image boundaries. While high-end computer vision methods on this task rely on sequence specific training of dedicated CNN architectures, we show the potential of a variational model, based on generic video information from motion and color. Such cues are usually required for tasks such as robot navigation or grasp estimation. We leverage them directly for video object segmentation and thus provide accurate segmentations at potentially very low extra cost. Our simple method can provide competitive results compared to the costly CNN-based methods with parameter tuning. Furthermore, we show that our approach can be combined with state-of-the-art CNN-based segmentations in order to improve over their respective results. We evaluate our method on the datasets DAVIS 16,17 and SegTrack v2.