Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

About vgg_v1.py #11

Open
koalaofpoint opened this issue May 22, 2019 · 1 comment
Open

About vgg_v1.py #11

koalaofpoint opened this issue May 22, 2019 · 1 comment

Comments

@koalaofpoint
Copy link

Sir,I'm refer your reseach to my work.I read your paper I found the earse means set the value=0; I need to know how to earse the feature,but the vgg1.py file I cann't understand.So I see maybe you can provide something for me,thank you

@xiaomengyc
Copy link
Owner

The feature maps to be erased are from the activation function ReLU, whose range is [0, +inf). Therefore, the actual erasing operation is to set the pixel values of object regions identified by Classifier A to zeros.

We erase the feature maps according to the right localization maps from classifier A in this line:
https://github.com/xiaomengyc/ACoL/blob/master/models/vgg/vgg_v1.py#L74.

Briefly, we construct a binary map, where pixels to be erased are 0, others are 1. Then, we multiply this map with the feature maps element-wisely. The piexls w.r.t the 0 positions are considered as erased.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants