You’re building a CNN to classify medical X-ray images (e.g., pneumonia vs normal). You apply Grad-CAM to understand which regions of an input image contribute most to the "pneumonia" class.
The Grad-CAM computation uses the last convolutional layer of the model:
What is the role of pooled_grads in the Grad-CAM explanation?

Question

You’re building a CNN to classify medical X-ray images (e.g., pneumonia vs normal). You apply Grad-CAM to understand which regions of an input image contribute most to the "pneumonia" class.

The Grad-CAM computation uses the last convolutional layer of the model:

What is the role of pooled_grads in the Grad-CAM explanation?

Answer

It computes the total activation for the final class

Answer

It indicates the model’s confidence in each pixel of the input image

Answer

It determines how much each channel in the feature map contributes to the predicted class

Answer

It is the final heatmap used for visualization

Crowdly

You’re building a CNN to classify medical X-ray images (e.g., pneumonia vs norm...

Want instant access to all verified answers on moodle.astanait.edu.kz?