arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.05917cs.CV
收藏

Point-VOS:指向式视频目标分割

Point-VOS: Pointing Up Video Object Segmentation

RWTH Aachen University(亚琛工业大学);Google Research(谷歌研究院)

Idil Esen Zulfikar, Sabarinath Mahadevan, Paul Voigtlaender, Bastian Leibe

AI总结:

本文提出Point-VOS任务,用时空稀疏点标注替代密集掩码标注以大幅降低视频目标分割的标注成本,构建了大规模点标注基准,并证明现有VOS方法基于伪掩码训练可接近全监督性能,同时该数据还能提升视频叙事定位任务。

摘要:

当前最先进的视频目标分割(VOS)方法在训练和测试阶段都依赖密集的逐目标掩码标注,这需要耗时且昂贵的视频标注机制。我们提出了一种新颖的Point-VOS任务,采用时空稀疏的点级标注方案,大幅降低了标注工作量。我们将该标注方案应用于两个带文本描述的大规模视频数据集,在32K个视频中标注了133K个目标上的超过19M个点。基于我们的标注,我们提出了一个新的Point-VOS基准,以及相应的基于点的训练机制,并以此建立了强基线结果。我们表明,现有的VOS方法可以轻松适配以在训练中利用我们的点标注,并且当在这些点生成的伪掩码上训练时,能够取得接近全监督性能的结果。此外,我们通过在视频叙事定位(VNG)任务上评估,表明我们的数据可用于改进连接视觉与语言的模型。我们将在https://pointvos.github.io公开代码和标注。

英文摘要:

Current state-of-the-art Video Object Segmentation (VOS) methods rely on dense per-object mask annotations both during training and testing. This requires time-consuming and costly video annotation mechanisms. We propose a novel Point-VOS task with a spatio-temporally sparse point-wise annotation scheme that substantially reduces the annotation effort. We apply our annotation scheme to two large-scale video datasets with text descriptions and annotate over 19M points across 133K objects in 32K videos. Based on our annotations, we propose a new Point-VOS benchmark, and a corresponding point-based training mechanism, which we use to establish strong baseline results. We show that existing VOS methods can easily be adapted to leverage our point annotations during training, and can achieve results close to the fully-supervised performance when trained on pseudo-masks generated from these points. In addition, we show that our data can be used to improve models that connect vision and language, by evaluating it on the Video Narrative Grounding (VNG) task. We will make our code and annotations available at https://pointvos.github.io.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑