UBoCo : Unsupervised Boundary Contrastive Learning for Generic Event Boundary Detection

Abstract

Generic Event Boundary Detection (GEBD) is a newly suggested video understanding task that aims to find one level deeper semantic boundaries of events. Bridging the gap between natural human perception and video understanding, it has various potential applications, including interpretable and semantically valid video parsing. Still at an early development stage, existing GEBD solvers are simple extensions of relevant video understanding tasks, disregarding GEBD's distinctive characteristics. In this paper, we propose a novel framework for unsupervised/supervised GEBD, by using the Temporal Self-similarity Matrix (TSM) as the video representation. The new Recursive TSM Parsing (RTP) algorithm exploits local diagonal patterns in TSM to detect boundaries, and it is combined with the Boundary Contrastive (BoCo) loss to train our encoder to generate more informative TSMs. Our framework can be applied to both unsupervised and supervised settings, with both achieving state-of-the-art performance by a huge margin in GEBD benchmark. Especially, our unsupervised method outperforms the previous state-of-the-art "supervised" model, implying its exceptional efficacy.

Motion Aware Self-Supervision for Generic Event Boundary Detection

“…On the Kinetics-GEBD dataset, our model outperforms the supervised baseline PC [68] with Rel. Dis threshold 0.05 and is also comparable with other stateof-the-art unsupervised/self-supervised GEBD models like UBoCo [36] and TeG [63] in terms of performance. Table 1 illustrates the result on the Kinetics-GEBD dataset.…”

“…Regarding unsupervised GEBD approaches, a shot detector library 2 and PredictAbility (PA) have been investigated in [68]. The authors of UBoCo [36,35] proposed a novel supervised/unsupervised method that applies contrastive learning to a TSM 3 based intermediary representation of videos to learn discriminatory boundary features. UBoCo's recursive TSM 3 parsing algorithm exploits generic patterns and detects very precise boundaries.…”