Local-Global Information Interaction Debiasing for Dynamic Scene Graph Generation

TL;DR

This paper introduces DynSGG-MTL, a multi-task learning model that enhances dynamic scene graph generation by integrating local and global information, effectively addressing long-tail predicate prediction issues in videos.

Contribution

It proposes a novel multi-task learning framework that combines local spatial-temporal and global human-action interactions for improved scene graph generation.

Findings

Improves accuracy of tail predicate predictions.

Alleviates long-tail distribution issues in scene graph datasets.

Enhances understanding of visual context in videos.

Abstract

The task of dynamic scene graph generation (DynSGG) aims to generate scene graphs for given videos, which involves modeling the spatial-temporal information in the video. However, due to the long-tailed distribution of samples in the dataset, previous DynSGG models fail to predict the tail predicates. We argue that this phenomenon is due to previous methods that only pay attention to the local spatial-temporal information and neglect the consistency of multiple frames. To solve this problem, we propose a novel DynSGG model based on multi-task learning, DynSGG-MTL, which introduces the local interaction information and global human-action interaction information. The interaction between objects and frame features makes the model more fully understand the visual context of the single image. Long-temporal human actions supervise the model to generate multiple scene graphs that conform to the global constraints and avoid the model being unable to learn the tail predicates. Extensive experiments on Action Genome dataset demonstrate the efficacy of our proposed framework, which not only improves the dynamic scene graph generation but also alleviates the long-tail problem.