Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval

TL;DR

This paper introduces a joint video and image encoder for end-to-end retrieval that leverages large-scale datasets and a curriculum learning approach, achieving state-of-the-art results in video-text retrieval tasks.

Contribution

The authors propose a novel adaptable model based on ViT and Timesformer architectures, trained with curriculum learning on image and video datasets, including a new WebVid-2M dataset.

Findings

Achieves state-of-the-art results on MSR-VTT, MSVD, DiDeMo, and LSMDC benchmarks.

Effective training on smaller datasets with curriculum learning.

Flexible model architecture for image and video text retrieval.

Abstract

Our objective in this work is video-text retrieval - in particular a joint embedding that enables efficient text-to-video retrieval. The challenges in this area include the design of the visual architecture and the nature of the training data, in that the available large scale video-text training datasets, such as HowTo100M, are noisy and hence competitive performance is achieved only at scale through large amounts of compute. We address both these challenges in this paper. We propose an end-to-end trainable model that is designed to take advantage of both large-scale image and video captioning datasets. Our model is an adaptation and extension of the recent ViT and Timesformer architectures, and consists of attention in both space and time. The model is flexible and can be trained on both image and video text datasets, either independently or in conjunction. It is trained with a curriculum learning schedule that begins by treating images as 'frozen' snapshots of video, and then gradually learns to attend to increasing temporal context when trained on video datasets. We also provide a new video-text pretraining dataset WebVid-2M, comprised of over two million videos with weak captions scraped from the internet. Despite training on datasets that are an order of magnitude smaller, we show that this approach yields state-of-the-art results on standard downstream video-retrieval benchmarks including MSR-VTT, MSVD, DiDeMo and LSMDC.