# Funnel-HOI: Top-Down Perception for Zero-Shot HOI Detection

arXiv: https://arxiv.org/abs/2507.12628 · HTML: https://mlsanity.com/paper/2507.12628/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 35 of its 64 references that are in the map.
- First posted on arXiv 16 Jul 2025.

## Builds on (most cited first, 30 of 35)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression](https://mlsanity.com/paper/1902.09630/) (2019)
- [SLIP: Self-supervision meets Language-Image Pre-training](https://mlsanity.com/paper/2112.12750/) (2021)
- [Learning to Detect Human-Object Interactions](https://mlsanity.com/paper/1702.05448/) (2017)
- [Detecting and Recognizing Human-Object Interactions](https://mlsanity.com/paper/1704.07333/) (2017)
- [Visual Semantic Role Labeling](https://mlsanity.com/paper/1505.04474/) (2015)
- [QPIC: Query-Based Pairwise Human-Object Interaction Detection with Image-Wide Contextual Information](https://mlsanity.com/paper/2103.05399/) (2021)
- [See More, Know More: Unsupervised Video Object Segmentation with Co-Attention Siamese Networks](https://mlsanity.com/paper/2001.06810/) (2020)
- [Visual Compositional Learning for Human-Object Interaction Detection](https://mlsanity.com/paper/2007.12407/) (2020)
- [iCAN: Instance-Centric Attention Network for Human-Object Interaction Detection](https://mlsanity.com/paper/1808.10437/) (2018)
- [GEN-VLKT: Simplify Association and Enhance Interaction Understanding for HOI Detection](https://mlsanity.com/paper/2203.13954/) (2022)
- [DRG: Dual Relation Graph for Human-Object Interaction Detection](https://mlsanity.com/paper/2008.11714/) (2020)
- [Mining the Benefits of Two-stage and One-stage HOI Detection](https://mlsanity.com/paper/2108.05077/) (2021)
- [Reformulating HOI Detection as Adaptive Set Prediction](https://mlsanity.com/paper/2103.05983/) (2021)
- [Spatially Conditioned Graphs for Detecting Human-Object Interactions](https://mlsanity.com/paper/2012.06060/) (2020)
- [Efficient Two-Stage Detection of Human-Object Interactions with a Novel Unary-Pairwise Transformer](https://mlsanity.com/paper/2112.01838/) (2021)
- [Detecting Human-Object Interaction via Fabricated Compositional Learning](https://mlsanity.com/paper/2103.08214/) (2021)
- [Affordance Transfer Learning for Human-Object Interaction Detection](https://mlsanity.com/paper/2104.02867/) (2021)
- [ALIP: Adaptive Language-Image Pre-training with Synthetic Caption](https://mlsanity.com/paper/2308.08428/) (2023)
- [HOICLIP: Efficient Knowledge Transfer for HOI Detection with Vision-Language Models](https://mlsanity.com/paper/2303.15786/) (2023)
- [Deep Contextual Attention for Human-Object Interaction Detection](https://mlsanity.com/paper/1910.07721/) (2019)
- [Detecting Human-Object Interactions via Functional Generalization](https://mlsanity.com/paper/1904.03181/) (2019)
- [ConsNet: Learning Consistency Graph for Zero-Shot Human-Object Interaction Detection](https://mlsanity.com/paper/2008.06254/) (2020)
- [Interact as You Intend: Intention-Driven Human-Object Interaction Detection](https://mlsanity.com/paper/1808.09796/) (2018)
- [End-to-End Zero-Shot HOI Detection via Vision and Language Knowledge Distillation](https://mlsanity.com/paper/2204.03541/) (2022)
- [FGAHOI: Fine-Grained Anchors for Human-Object Interaction Detection](https://mlsanity.com/paper/2301.04019/) (2023)
- [Discovering Human-Object Interaction Concepts via Self-Compositional Learning](https://mlsanity.com/paper/2203.14272/) (2022)

## Abstract (arXiv)

Human-object interaction detection (HOID) refers to localizing interactive human-object pairs in images and identifying the interactions. Since there could be an exponential number of object-action combinations, labeled data is limited - leading to a long-tail distribution problem. Recently, zero-shot learning emerged as a solution, with end-to-end transformer-based object detectors adapted for HOID becoming successful frameworks. However, their primary focus is designing improved decoders for learning entangled or disentangled interpretations of interactions. We advocate that HOI-specific cues must be anticipated at the encoder stage itself to obtain a stronger scene interpretation. Consequently, we build a top-down framework named Funnel-HOI inspired by the human tendency to grasp well-defined concepts first and then associate them with abstract concepts during scene understanding. We first probe an image for the presence of objects (well-defined concepts) and then probe for actions (abstract concepts) associated with them. A novel asymmetric co-attention mechanism mines these cues utilizing multimodal information (incorporating zero-shot capabilities) and yields stronger interaction representations at the encoder level. Furthermore, a novel loss is devised that considers objectaction relatedness and regulates misclassification penalty better than existing loss functions for guiding the interaction classifier. Extensive experiments on the HICO-DET and V-COCO datasets across fully-supervised and six zero-shot settings reveal our state-of-the-art performance, with up to 12.4% and 8.4% gains for unseen and rare HOI categories, respectively.