Code for the paper CNN fixations: An unravelling approach to identify discriminative image regions

Konda Reddy Mopuri, Utsav Garg, R. Venkatesh Babu

This repository can be used to visualize predictions for four CNNs namely: AlexNet, VGG-16, GoogLeNet and ResNet-101.

1. In the demo folder add path to caffe installation

2. Install the required dependencies

3. Navigate to the Demo folder and open a ipython notebook

Add the .caffemodels for the respective architechtures before running the codes

Results of running the demo for one image from ILSVRC validation set

More Samples

The VGG-16 demo now also has the option to use the TensorFlow backend in addition to the default Caffe backend. (Contributed by Joe Yearsley)

Contact Utsav Garg if you have questions.

CNN Fixations: An unraveling approach to visualize the discriminative image regions