Diagnosing Rarity in Human-Object Interaction Detection

CVPRW 2020 pp. 3956-3960

doi:10.1109/CVPRW50498.2020.00460 /cvprw/2020/kilickaya2020cvprw-diagnosing/

Abstract

Human-object interaction (HOI) detection is a core task in computer vision. The goal is to localize all human-object pairs and recognize their interactions. An interaction de-fined by a tuple leads to a long-tailed visual recognition challenge since many combinations are rarely represented. The performance of the proposed models is limited especially for the tail categories, but little has been done to understand the reason. To that end, in this paper, we propose to diagnose rarity in HOI detection. We propose a three-step strategy, namely Detection, Identification and Recognition where we carefully analyse the limiting factors by studying state-of-the-art models. Our findings indicate that detection and identification steps are altered by the interaction signals like occlusion and relative location, as a result limiting the recognition accuracy.

PDF CVPRW Semantic Scholar

Cite

Text

Kilickaya and Smeulders. "Diagnosing Rarity in Human-Object Interaction Detection." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020. doi:10.1109/CVPRW50498.2020.00460

Markdown

[Kilickaya and Smeulders. "Diagnosing Rarity in Human-Object Interaction Detection." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020.](https://mlanthology.org/cvprw/2020/kilickaya2020cvprw-diagnosing/) doi:10.1109/CVPRW50498.2020.00460

BibTeX

@inproceedings{kilickaya2020cvprw-diagnosing,
  title     = {{Diagnosing Rarity in Human-Object Interaction Detection}},
  author    = {Kilickaya, Mert and Smeulders, Arnold W. M.},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2020},
  pages     = {3956-3960},
  doi       = {10.1109/CVPRW50498.2020.00460},
  url       = {https://mlanthology.org/cvprw/2020/kilickaya2020cvprw-diagnosing/}
}