search

yuantn / MI-AOD

Code for Multiple Instance Active Learning for Object Detection, CVPR 2021
https://openaccess.thecvf.com/content/CVPR2021/papers/Yuan_Multiple_Instance_Active_Learning_for_Object_Detection_CVPR_2021_paper.pdf
Apache License 2.0
333 stars 43 forks source link

Very basic doubt about definition of a symbol #78

Closed skbhat closed 2 years ago

skbhat commented 2 years ago

Namaste,

I would like to know how $y_c^{cls}$ is computed for an image. It is not described in the paper. I assume that if an object belonging to category $c$ exists in the image then $y_c^{cls}=1$, else $y_c^{cls}=0$. Is that correct?

thanks and regards Srikrishna.

yuantn commented 2 years ago

Namaste,

$y^{cls}_c$ is just what you assumed, and it is from the dataset annotation. For example, if there is some objects that belong to the 3-rd class and the 5-th class. The original annotation may be [2,4], and the corresponding $y^{cls}$ is a one-hot vector [0,0,1,0,1].

Best regards, Tianning Yuan.