Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

RLIP: Relational Language-Image Pre-training for Human-Object Interaction Detection

About

The task of Human-Object Interaction (HOI) detection targets fine-grained visual parsing of humans interacting with their environment, enabling a broad range of applications. Prior work has demonstrated the benefits of effective architecture design and integration of relevant cues for more accurate HOI detection. However, the design of an appropriate pre-training strategy for this task remains underexplored by existing approaches. To address this gap, we propose Relational Language-Image Pre-training (RLIP), a strategy for contrastive pre-training that leverages both entity and relation descriptions. To make effective use of such pre-training, we make three technical contributions: (1) a new Parallel entity detection and Sequential relation inference (ParSe) architecture that enables the use of both entity and relation descriptions during holistically optimized pre-training; (2) a synthetic data generation framework, Label Sequence Extension, that expands the scale of language data available within each minibatch; (3) mechanisms to account for ambiguity, Relation Quality Labels and Relation Pseudo-Labels, to mitigate the influence of ambiguous/noisy samples in the pre-training data. Through extensive experiments, we demonstrate the benefits of these contributions, collectively termed RLIP-ParSe, for improved zero-shot, few-shot and fine-tuning HOI detection performance as well as increased robustness to learning from noisy annotations. Code will be available at https://github.com/JacobYuan7/RLIP.

Hangjie Yuan, Jianwen Jiang, Samuel Albanie, Tao Feng, Ziyuan Huang, Dong Ni, Mingqian Tang• 2022

Related benchmarks

TaskDatasetResultRank
Human-Object Interaction DetectionHICO-DET (test)
mAP (full)32.84
493
Human-Object Interaction DetectionV-COCO (test)
AP (Role, Scenario 1)61.9
270
Human-Object Interaction DetectionHICO-DET
mAP (Full)32.84
233
Human-Object Interaction DetectionHICO-DET (Rare First Unseen Combination (RF-UC))
mAP (Full)30.52
77
Human-Object Interaction DetectionV-COCO 1.0 (test)
AP_role (#1)61.9
76
Human-Object Interaction DetectionV-COCO
AP^1 Role61.9
65
HOI DetectionV-COCO
AP Role 161.9
40
Human-Object Interaction DetectionHICO-DET (NF-UC)
mAP (Full)26.19
40
HOI DetectionHICO-DET
mAP (Rare)26.85
34
Human-Object Interaction DetectionHICO-DET Zero-Shot
mAP (Default Unseen)20.27
33
Showing 10 of 24 rows

Other info

Code

Follow for update