arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1708.02696cs.CV
收藏

了解视频中的人类行为需要采取哪些行动?

What Actions are Needed for Understanding Human Actions in Videos?

Carnegie Mellon University(卡内基梅隆大学)

Gunnar A. Sigurdsson, Olga Russakovsky, Abhinav Gupta

摘要:

对人类活动进行推理的正确方法是什么?哪些前进方向最有希望?在这项工作中,我们分析了视频中人类活动理解的现状。本文的目标是检查数据集,评估指标,算法和潜在的未来方向。我们着眼于定义活动的定性属性,如姿势可变性,简洁性和密度。实验考虑了多个最先进的算法和多个数据集。结果表明,虽然有内在的模糊性的时间范围内的活动,目前的数据集仍然允许有效的基准。我们发现,细粒度的理解对象和姿势结合时间推理时,可能会产生实质性的改进算法的准确性。我们提出了许多类型的信息,将需要实现实质性的收益,在活动的理解:对象,动词,意图和顺序推理。该软件和其他信息将提供给其他研究人员,以了解他们自己的算法。

英文摘要:

What is the right way to reason about human activities? What directions forward are most promising? In this work, we analyze the current state of human activity understanding in videos. The goal of this paper is to examine datasets, evaluation metrics, algorithms, and potential future directions. We look at the qualitative attributes that define activities such as pose variability, brevity, and density. The experiments consider multiple state-of-the-art algorithms and multiple datasets. The results demonstrate that while there is inherent ambiguity in the temporal extent of activities, current datasets still permit effective benchmarking. We discover that fine-grained understanding of objects and pose when combined with temporal reasoning is likely to yield substantial improvements in algorithmic accuracy. We present the many kinds of information that will be needed to achieve substantial gains in activity understanding: objects, verbs, intent, and sequential reasoning. The software and additional information will be made available to provide other researchers detailed diagnostics to understand their own algorithms.

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。

↑