HowTo100M: Learning a Text-Video Embedding by Watching Hundred Million Narrated Video Clips
This work addresses the scalability issue in creating large-scale text-video datasets for researchers and practitioners in computer vision and multimedia, though it is incremental as it builds on existing embedding methods.
The authors tackled the problem of learning text-video embeddings without expensive manual captions by using automatically transcribed narrations from instructional videos, resulting in state-of-the-art performance on text-to-video retrieval and action localization tasks, with improvements shown on datasets like YouCook2 and CrossTask.
Learning text-video embeddings usually requires a dataset of video clips with manually provided captions. However, such datasets are expensive and time consuming to create and therefore difficult to obtain on a large scale. In this work, we propose instead to learn such embeddings from video data with readily available natural language annotations in the form of automatically transcribed narrations. The contributions of this work are three-fold. First, we introduce HowTo100M: a large-scale dataset of 136 million video clips sourced from 1.22M narrated instructional web videos depicting humans performing and describing over 23k different visual tasks. Our data collection procedure is fast, scalable and does not require any additional manual annotation. Second, we demonstrate that a text-video embedding trained on this data leads to state-of-the-art results for text-to-video retrieval and action localization on instructional video datasets such as YouCook2 or CrossTask. Finally, we show that this embedding transfers well to other domains: fine-tuning on generic Youtube videos (MSR-VTT dataset) and movies (LSMDC dataset) outperforms models trained on these datasets alone. Our dataset, code and models will be publicly available at: www.di.ens.fr/willow/research/howto100m/.