Boosting Human-Object Interaction Detection with Text-to-Image Diffusion Model

TL;DR

This paper introduces DiffHOI, a novel HOI detection method leveraging a pre-trained text-to-image diffusion model and a synthetic dataset SynHOI, significantly improving detection accuracy and addressing data scarcity issues.

Contribution

The paper presents a new HOI detection framework using a frozen diffusion model for enhanced representations and a large-scale synthetic dataset to mitigate data imbalance.

Findings

DiffHOI achieves 41.50 mAP, outperforming state-of-the-art methods.

SynHOI reduces long-tail issues and boosts rare class detection by 11.55% mAP.

The approach improves zero-shot and model-agnostic HOI detection performance.

Abstract

This paper investigates the problem of the current HOI detection methods and introduces DiffHOI, a novel HOI detection scheme grounded on a pre-trained text-image diffusion model, which enhances the detector's performance via improved data diversity and HOI representation. We demonstrate that the internal representation space of a frozen text-to-image diffusion model is highly relevant to verb concepts and their corresponding context. Accordingly, we propose an adapter-style tuning method to extract the various semantic associated representation from a frozen diffusion model and CLIP model to enhance the human and object representations from the pre-trained detector, further reducing the ambiguity in interaction prediction. Moreover, to fill in the gaps of HOI datasets, we propose SynHOI, a class-balance, large-scale, and high-diversity synthetic dataset containing over 140K HOI images with fully triplet annotations. It is built using an automatic and scalable pipeline designed to scale up the generation of diverse and high-precision HOI-annotated data. SynHOI could effectively relieve the long-tail issue in existing datasets and facilitate learning interaction representations. Extensive experiments demonstrate that DiffHOI significantly outperforms the state-of-the-art in regular detection (i.e., 41.50 mAP) and zero-shot detection. Furthermore, SynHOI can improve the performance of model-agnostic and backbone-agnostic HOI detection, particularly exhibiting an outstanding 11.55% mAP improvement in rare classes.