Object-aware Video-language Pre-training for Retrieval

TL;DR

This paper introduces Object-aware Transformers, a novel video-language pre-training model that explicitly incorporates object representations to improve semantic alignment and retrieval performance across multiple benchmarks.

Contribution

It presents an object-centric extension to video-language transformers that leverages object bounding boxes and tags for enhanced semantic understanding.

Findings

Consistent performance improvements across all evaluated benchmarks.

Effective integration of object information enhances semantic alignment.

Deep analysis and ablation studies validate the proposed approach.

Abstract

Recently, by introducing large-scale dataset and strong transformer network, video-language pre-training has shown great success especially for retrieval. Yet, existing video-language transformer models do not explicitly fine-grained semantic align. In this work, we present Object-aware Transformers, an object-centric approach that extends video-language transformer to incorporate object representations. The key idea is to leverage the bounding boxes and object tags to guide the training process. We evaluate our model on three standard sub-tasks of video-text matching on four widely used benchmarks. We also provide deep analysis and detailed ablation about the proposed method. We show clear improvement in performance across all tasks and datasets considered, demonstrating the value of a model that incorporates object representations into a video-language architecture. The code will be released at \url{https://github.com/FingerRec/OA-Transformer}.