490.490. Grad-CAM

Grad-CAM

In this problem, you will be implementing a function to generate Grad-CAM heatmaps from a given input image and a pre-trained model. Grad-CAM is a technique 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 Grad-CAM 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 Grad-CAM 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.