What Actions are Needed for Understanding Human Actions in Videos?

TLDR

Human activity understanding demands principled reasoning and promising research directions. The paper reviews datasets, metrics, algorithms, and future directions in human activity understanding, and identifies objects, verbs, intent, and sequential reasoning as key information for substantial gains. The authors examine qualitative activity attributes, benchmark multiple state‑of‑the‑art algorithms across datasets, and release software and diagnostics for community use. Results show that despite temporal ambiguity, current datasets enable effective benchmarking, and that fine‑grained object and pose understanding combined with temporal reasoning can substantially improve algorithmic accuracy.

Abstract

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.