Additive Class Distinction Maps using Branched-GANs
This work addresses the need for explainable AI and self-supervised segmentation in domains such as medical imaging and computer vision, though it is presented as a preliminary report.
The authors tackled the problem of creating precise pixel-wise distinction maps between two image classes using a branched-GAN architecture based on image decomposition, resulting in interpretable visualizations applicable to tasks like MRI tumor extraction and face feature isolation.
We present a new model, training procedure and architecture to create precise maps of distinction between two classes of images. The objective is to comprehend, in pixel-wise resolution, the unique characteristics of a class. These maps can facilitate self-supervised segmentation and objectdetection in addition to new capabilities in explainable AI (XAI). Our proposed architecture is based on image decomposition, where the output is the sum of multiple generative networks (branched-GANs). The distinction between classes is isolated in a dedicated branch. This approach allows clear, precise and interpretable visualization of the unique characteristics of each class. We show how our generic method can be used in several modalities for various tasks, such as MRI brain tumor extraction, isolating cars in aerial photography and obtaining feminine and masculine face features. This is a preliminary report of our initial findings and results.