Visual Feature Attribution using Wasserstein GANs

TL;DR

This paper introduces a novel visual feature attribution method using Wasserstein GANs that overcomes limitations of existing techniques, demonstrating superior performance on synthetic and neuroimaging datasets, including realistic disease effect maps.

Contribution

The paper proposes a new feature attribution technique based on Wasserstein GANs that addresses limitations of current methods, improving interpretability in medical imaging.

Findings

Outperforms state-of-the-art methods on synthetic data

Produces realistic disease effect maps in neuroimaging data

Effective in identifying category-specific features

Abstract

Attributing the pixels of an input image to a certain category is an important and well-studied problem in computer vision, with applications ranging from weakly supervised localisation to understanding hidden effects in the data. In recent years, approaches based on interpreting a previously trained neural network classifier have become the de facto state-of-the-art and are commonly used on medical as well as natural image datasets. In this paper, we discuss a limitation of these approaches which may lead to only a subset of the category specific features being detected. To address this problem we develop a novel feature attribution technique based on Wasserstein Generative Adversarial Networks (WGAN), which does not suffer from this limitation. We show that our proposed method performs substantially better than the state-of-the-art for visual attribution on a synthetic dataset and on real 3D neuroimaging data from patients with mild cognitive impairment (MCI) and Alzheimer's disease (AD). For AD patients the method produces compellingly realistic disease effect maps which are very close to the observed effects.