Real Time Image Saliency for Black Box Classifiers

TL;DR

This paper introduces a fast, real-time saliency detection method applicable to any differentiable image classifier, producing interpretable and artifact-free saliency maps suitable for practical use.

Contribution

The authors develop a universal, efficient saliency detection model that generalizes well and outperforms existing weakly supervised methods on large-scale datasets.

Findings

Achieves real-time saliency detection with a single forward pass

Produces sharp, artifact-free, and interpretable saliency maps

Outperforms other weakly supervised methods on ImageNet localization

Abstract

In this work we develop a fast saliency detection method that can be applied to any differentiable image classifier. We train a masking model to manipulate the scores of the classifier by masking salient parts of the input image. Our model generalises well to unseen images and requires a single forward pass to perform saliency detection, therefore suitable for use in real-time systems. We test our approach on CIFAR-10 and ImageNet datasets and show that the produced saliency maps are easily interpretable, sharp, and free of artifacts. We suggest a new metric for saliency and test our method on the ImageNet object localisation task. We achieve results outperforming other weakly supervised methods.