PhD Seminar • Artificial Intelligence/Computer Vision — Attention Mechanisms for Weakly Supervised Image Segmentation

Paulo Pacheco, PhD candidate
David R. Cheriton School of Computer Science

In this seminar I will discuss some of the current state-of-the-art methods for generating attention maps in weakly supervised (image level annotations) for classification tasks, aiming image segmentation inference and object localization.

I will present techniques that use the Global Average Polling Layer, Backpropagation and Image Masking, presenting some experimental results that compare these techniques under leveraged conditions.