Awesome

HICO-DET

Utilities for the human-object interaction detection dataset HICO-DET

Supported Utilities

• NEW! Train and test advanced variants of DETR on HICO-DET
• Train and test DETR on HICO-DET
• A command-line style dataset navigator
• Large-scale visualisation in web page
• Generate object detections with Faster R-CNN
• Generate ground truth object detections
• Visualise detected objects
• Evaluate object detections
• Fine-tune Faster R-CNN on HICO-DET

Installation Instructions

1. Download the repo with git clone https://github.com/fredzzhang/hicodet.git.
2. Prepare the HICO-DET dataset.
   1. If you have not downloaded the dataset before, run the following script.

   cd /path/to/hicodet
   bash download.sh
   

   2. If you have previously downloaded the dataset, simply create a soft link.

   cd /path/to/hicodet
   ln -s /path/to/hico_20160224_det ./hico_20160224_det
   

3. Install the lightweight deep learning library Pocket if you haven't yet.
4. Make sure the environment you created for Pocket is activated. You are good to go!

Citation

If you find our work useful for your research, please consider citing us:

@inproceedings{zhang2023pvic,
  author    = {Zhang, Frederic Z. and Yuan, Yuhui and Campbell, Dylan and Zhong, Zhuoyao and Gould, Stephen},
  title     = {Exploring Predicate Visual Context in Detecting Human–Object Interactions},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  month     = {October},
  year      = {2023},
  pages     = {10411-10421},
}

@inproceedings{zhang2022upt,
  author    = {Zhang, Frederic Z. and Campbell, Dylan and Gould, Stephen},
  title     = {Efficient Two-Stage Detection of Human-Object Interactions with a Novel Unary-Pairwise Transformer},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  month     = {June},
  year      = {2022},
  pages     = {20104-20112}
}

@inproceedings{zhang2021scg,
  author    = {Zhang, Frederic Z. and Campbell, Dylan and Gould, Stephen},
  title     = {Spatially Conditioned Graphs for Detecting Human–Object Interactions},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  month     = {October},
  year      = {2021},
  pages     = {13319-13327}
}

Dataset Class

The implementation of the dataset class can be found in hicodet.py. Refer to the documentation to find out more about its usage. For convenience, the dataset class has been included in the Pocket library, accessible via pocket.data.HICODet.

License

MIT License