The official repository for Deformable ProtoPNet, as described in "Deformable ProtoPNet: An Interpretable Image Classifier Using Deformable Prototypes".
MIT License
37
stars
11
forks
source link
Get bounding boxes of prototypes for each image #3
I was able to successfully train a Deformable ProtoPNet following the README instructions.
I had a question a question because I am not sure how to implement it myself. What would be the easiest way to make a script that would compute and save for each image:
the activation map of each prototype (un-upsampled, and upsampled)
the highest activation crop of each prototype (bounding box)
I would be very grateful if you find the time to help me on this.
Thanks :)
Elias
here is my email address: elias.ramzi@lecnam.net
Hi,
I was able to successfully train a Deformable ProtoPNet following the README instructions.
I had a question a question because I am not sure how to implement it myself. What would be the easiest way to make a script that would compute and save for each image:
I would be very grateful if you find the time to help me on this.
Thanks :) Elias here is my email address: elias.ramzi@lecnam.net