Multimodal Pretraining for Dense Video Captioning
This work addresses the need for improved user experience in learning hands-on skills via instructional videos, though it is incremental as it builds on existing dense video captioning methods.
The authors tackled the problem of automatically generating time-stamped annotations for instructional videos by constructing a new dataset (ViTT) and exploring multimodal pretraining strategies, resulting in models that generalize well across a variety of videos.
Learning specific hands-on skills such as cooking, car maintenance, and home repairs increasingly happens via instructional videos. The user experience with such videos is known to be improved by meta-information such as time-stamped annotations for the main steps involved. Generating such annotations automatically is challenging, and we describe here two relevant contributions. First, we construct and release a new dense video captioning dataset, Video Timeline Tags (ViTT), featuring a variety of instructional videos together with time-stamped annotations. Second, we explore several multimodal sequence-to-sequence pretraining strategies that leverage large unsupervised datasets of videos and caption-like texts. We pretrain and subsequently finetune dense video captioning models using both YouCook2 and ViTT. We show that such models generalize well and are robust over a wide variety of instructional videos.