arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2307.09356cs.CV
收藏

OnlineRefer:一种用于指代视频目标分割的简单在线基线

OnlineRefer: A Simple Online Baseline for Referring Video Object Segmentation

MEGVII Technology(旷视科技);Shanghai Jiao Tong University(上海交通大学);Beijing Academy of Artificial Intelligence(北京智源人工智能研究院);Beijing Institute of Technology(北京理工大学);University of Macau(澳门大学)

Dongming Wu, Tiancai Wang, Yuang Zhang, Xiangyu Zhang, Jianbing Shen

AI总结:

针对现有指代视频目标分割离线方法时序关联建模有限的问题,提出带显式查询传播的在线模型OnlineRefer及半在线框架,在四个基准上性能超越现有离线方法。

摘要:

指代视频目标分割(RVOS)旨在根据人类指令分割视频中的目标。当前最先进的方法都属于离线模式,在该模式下每个视频片段独立与文本嵌入交互以实现跨模态理解。这类方法通常认为离线模式对RVOS是必要的,但仅能在单个片段内建模有限的时序关联。本研究打破了此前的离线认知,提出了一种采用显式查询传播的简单且有效的在线模型,命名为OnlineRefer。具体而言,该方法利用汇聚了语义信息和位置先验的目标线索,提升当前帧指代预测的准确性和便捷性。此外,我们将该在线模型推广为半在线框架,以适配基于视频的骨干网络。为验证方法的有效性,我们在四个基准数据集上进行了评估,即Refer-Youtube-VOS、Refer-DAVIS17、A2D-Sentences和JHMDB-Sentences。在未使用额外技巧的情况下,采用Swin-L骨干网络的OnlineRefer在Refer-Youtube-VOS和Refer-DAVIS17上分别取得63.5 J&F和64.8 J&F的成绩,性能超过所有其他离线方法。

英文摘要:

Referring video object segmentation (RVOS) aims at segmenting an object in a video following human instruction. Current state-of-the-art methods fall into an offline pattern, in which each clip independently interacts with text embedding for cross-modal understanding. They usually present that the offline pattern is necessary for RVOS, yet model limited temporal association within each clip. In this work, we break up the previous offline belief and propose a simple yet effective online model using explicit query propagation, named OnlineRefer. Specifically, our approach leverages target cues that gather semantic information and position prior to improve the accuracy and ease of referring predictions for the current frame. Furthermore, we generalize our online model into a semi-online framework to be compatible with video-based backbones. To show the effectiveness of our method, we evaluate it on four benchmarks, \ie, Refer-Youtube-VOS, Refer-DAVIS17, A2D-Sentences, and JHMDB-Sentences. Without bells and whistles, our OnlineRefer with a Swin-L backbone achieves 63.5 J&F and 64.8 J&F on Refer-Youtube-VOS and Refer-DAVIS17, outperforming all other offline methods.

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。

↑