Learning Temporal Sentence Grounding From Narrated EgoVideos

TL;DR

This paper introduces CliMer, a contrastive learning approach that uses narrations and rough timestamps to improve temporal sentence grounding in egocentric videos, especially on longer, fine-grained datasets.

Contribution

It presents a novel clip merging technique for training TSG models using only narration data, enhancing performance on egocentric datasets.

Findings

Mean R@1 improves from 3.9 to 5.7 on Ego4D.

Mean R@1 improves from 10.7 to 13.0 on EPIC-Kitchens.

Effective contrastive training with clip merging enhances TSG accuracy.

Abstract

The onset of long-form egocentric datasets such as Ego4D and EPIC-Kitchens presents a new challenge for the task of Temporal Sentence Grounding (TSG). Compared to traditional benchmarks on which this task is evaluated, these datasets offer finer-grained sentences to ground in notably longer videos. In this paper, we develop an approach for learning to ground sentences in these datasets using only narrations and their corresponding rough narration timestamps. We propose to artificially merge clips to train for temporal grounding in a contrastive manner using text-conditioning attention. This Clip Merging (CliMer) approach is shown to be effective when compared with a high performing TSG method -- e.g. mean R@1 improves from 3.9 to 5.7 on Ego4D and from 10.7 to 13.0 on EPIC-Kitchens. Code and data splits available from: https://github.com/keflanagan/CliMer