491.491. Saliency Maps

Saliency Maps

In this problem, you will be implementing a function to generate saliency maps from a given input image and a pre-trained model. Saliency maps are used to visualize the importance of different regions of the input image for a specific class. Your function should take in an input image, a pre-trained model, and a class index, and return a heatmap representation of the saliency scores.

Example:

Suppose we have an input image x, a pre-trained model model, and a class index class_idx. The function should return a heatmap with the saliency scores plotted.

Constraints:

* The input image should be a 3D numpy array with shape (height, width, channels).

* The pre-trained model should be a Keras model.

* The class index should be an integer.