search

AutoNUE / public-code

http://cvit.iiit.ac.in/scene-understanding-challenge-2018/
105 stars 49 forks source link

Sample input and output of instance segmentation evaluation script #6

Closed kHarshit closed 4 years ago

kHarshit commented 5 years ago

The evaluate_instance_segmentation.py takes input defined as follows:

https://github.com/AutoNUE/public-code/blob/b2d44a4b7b05b66b05af68dbc35af058235f24ea/evaluation/evaluate_instance_segmentation.py#L13-L29

Do multiple files need to be created (one for each label) per image?

Can you provide such sample file for validation images in the dataset so as to check the working of the script. Or for at least one image? Or tell me how do I generate such file from *_gtFine_instanceids.png?

anirudh-chakravarthy commented 5 years ago

Hi,

I was facing the same issue. For instance segmentation, I am unable to create text files containing the predictions of the given format. Were you able to generate text files?

kHarshit commented 5 years ago

Hi @anirudh-chakravarthy ,

AFAIK, you need to create separate binary mask file for each predicted instance. You can save those masks from the model you're using.

car