This repository helps to create a saliency map of an image using ResNet50 framework, trained on ImageNet dataset. First execute the function.py file, it will load the image from the wikimedia, you can also import your own image from local disk by using simler function. Then execute the test.py file. It import the ResNet50 from keras API. Then execute pred.py, it will predict the outcomes of your model with higher probability. At the end execute saliency_map.py. It will create a saliecy map of an image.
Input image
(1)Saliency_map
(2)Input_image
(3)Overlayed_image