search

OreoChocolate / MUREN

The official code for Relational Context Learning for Human-Object Interaction Detection, CVPR2023.
http://cvlab.postech.ac.kr/research/MUREN/
48 stars 5 forks source link

category_id #8

Closed anjugopinath closed 7 months ago

anjugopinath commented 7 months ago

Hi,

Why are there two sets of category ids as shown in the red and blue boxes? From where can I get what this category id maps to?

image

Thank You.

edugzlez commented 7 months ago

category_id is the category of the COCO dataset

https://github.com/amikelive/coco-labels/blob/master/coco-labels-paper.txt

OreoChocolate commented 7 months ago

First of all, thank you for your interest in our research.

In response, our code is built upon the codebase of CDN (Zhang et al., NeurIPS 2021), so we are using the annotations utilized in CDN. If you visit the CDN GitHub page, you can see that they convert the annotations (i.e., category_id) from HICO-DET and V-COCO into the format proposed by PPDM (Liao et al., CVPR 2020).

Additionally, we are sharing the reference list for annotation index through the following link: https://gist.github.com/OreoChocolate/85721bcadf943f54c3432369a1ccf414

The object index for V-COCO is the same as the object index for HICO-DET.

Thank you.

anjugopinath commented 7 months ago

@edugzlez and @OreoChocolate , Thank You for your response. This clarifies my doubt. I was confused earlier, since I was referring the mappings given on the HICO-DET website:

https://websites.umich.edu/~ywchao/hico/

image

but the ids did not add up when I checked one of the images in HICO-DET and the corresponding annotations.