CEFA: A Plug-and-Play Method for Rare Human-Object Interactions Detection by Bridging Domain Gap

Code for our ACM MM24 paper "A Plug-and-Play Method for Rare Human-Object Interactions Detection by Bridging Domain Gap" .

Contributed by Lijun Zhang, Wei Suo, Peng Wang, Yanning Zhang.

Installation

Install the dependencies.

pip install -r requirements.txt

Data preparation

HICO-DET

HICO-DET dataset can be downloaded here. After finishing downloading, unpack the tarball (hico_20160224_det.tar.gz) to the data directory.

V-COCO

First clone the repository of V-COCO from here, and then follow the instruction to generate the file instances_vcoco_all_2014.json. Next, download the prior file prior.pickle from here.

Training

HICO-DET

# default setting
sh ./scripts/train.sh

Acknowledge

Codes are built from GEN-VLKT, HOICLIP , DETR, QPIC and CDN. We thank them for their contributions.

Release Schedule

• We will update more detailed README.md (including dataset, training, verification) in the future