Weakly Supervised Dense Video Captioning via Jointly Usage of Knowledge Distillation and Cross-modal Matching

TL;DR

This paper introduces a weakly supervised dense video captioning method that leverages knowledge distillation and cross-modal matching to generate high-quality captions without pairwise annotations, outperforming existing methods.

Contribution

It presents a novel approach combining knowledge distillation and cross-modal retrieval for weakly supervised dense video captioning, with improved proposal generation and semantic matching.

Findings

Outperforms state-of-the-art methods on ActivityNet-Caption dataset.

Effective use of knowledge distillation improves event proposal quality.

Cross-modal matching enhances semantic alignment between proposals and sentences.

Abstract

This paper proposes an approach to Dense Video Captioning (DVC) without pairwise event-sentence annotation. First, we adopt the knowledge distilled from relevant and well solved tasks to generate high-quality event proposals. Then we incorporate contrastive loss and cycle-consistency loss typically applied to cross-modal retrieval tasks to build semantic matching between the proposals and sentences, which are eventually used to train the caption generation module. In addition, the parameters of matching module are initialized via pre-training based on annotated images to improve the matching performance. Extensive experiments on ActivityNet-Caption dataset reveal the significance of distillation-based event proposal generation and cross-modal retrieval-based semantic matching to weakly supervised DVC, and demonstrate the superiority of our method to existing state-of-the-art methods.