search

erikalu / class-agnostic-counting

63 stars 14 forks source link

Reading Dot annotations #2

Closed akshitac8 closed 5 years ago

akshitac8 commented 5 years ago

Hey @erikalu I am unable to understand the line

lb = np.asarray(lb[:, :, 0] > 230)

I believe that this takes all the values greater than 230 in the red channel. However, dot annotations are a binary mask of shape ((Height, Width)).

It would be great if you could explain the use case for this line.

Thanks, Akshita Gupta

WeidiXie commented 5 years ago

have you checked the dot annotation map ? how do you know it's binary mask ?

erikalu commented 5 years ago

Hi @akshitac8 you are correct, this line of code was for a specific dataset where the labels were in this form (red channel). You can modify it to suit your needs.

akshitac8 commented 5 years ago

Thank you @erikalu :)