Optimized Global Perturbation Attacks For Brain Tumour ROI Extraction From Binary Classification Models

TL;DR

This paper introduces an optimized perturbation attack method to extract brain tumour regions of interest from binary classification models, addressing challenges in medical image segmentation with limited annotated data.

Contribution

It presents a novel weakly supervised approach and a new objective function for ROI extraction in brain tumour MRI using binary labels and pretrained classifiers.

Findings

Effective ROI extraction from binary classifiers

Improved segmentation accuracy in MRI brain tumour data

Reduced need for detailed annotations

Abstract

Deep learning techniques have greatly benefited computer-aided diagnostic systems. However, unlike other fields, in medical imaging, acquiring large fine-grained annotated datasets such as 3D tumour segmentation is challenging due to the high cost of manual annotation and privacy regulations. This has given interest to weakly-supervise methods to utilize the weakly labelled data for tumour segmentation. In this work, we propose a weakly supervised approach to obtain regions of interest using binary class labels. Furthermore, we propose a novel objective function to train the generator model based on a pretrained binary classification model. Finally, we apply our method to the brain tumour segmentation problem in MRI.