TAAFT
Free mode
100% free
Freemium
Free Trial
Create tool

Grad-CAM

[græd kæm]
Computer Vision
Last updated: December 9, 2024

Definition

A technique that creates heat maps showing which regions of an image are important for classification.

Detailed Explanation

Gradient-weighted Class Activation Mapping uses the gradients of any target concept flowing into the final convolutional layer to produce a coarse localization map highlighting important regions in the image.

Use Cases

Medical diagnosis visualization, autonomous vehicle decision explanation, quality control inspection

Related Terms