The Instantaneous Accuracy: a Novel Metric for the Problem of Online Human Behaviour Recognition in Untrimmed Videos

The problem of Online Human Behavior Recognition in untrimmed videos, aka Online Action Detection (OAD), needs to be revisited. Unlike traditional off-line action detection approaches, where the evaluation metrics are clear and well established, in the OAD setting we find few works and no consensus on evaluation protocols to be used. In this paper we introduce a novel online metric, the Instantaneous Accuracy (IA), that exhibits an online nature, solving most of the limitations of the previous (off-line) metrics. We conduct a thorough experimental evaluation on the TVSeries dataset, comparing the performance of various baseline methods with the state of the art. Our results confirm the problems of the previous evaluation protocols, and suggest that an IA-based protocol is more adequate to the online scenario for human behaviour understanding.