Deep Superpixel Generation and Clustering for Weakly Supervised Segmentation of Brain Tumors in MR Images

TL;DR

This paper introduces a superpixel-based weakly supervised segmentation method for brain tumors in MR images, leveraging image-level labels to improve segmentation accuracy without requiring detailed annotations.

Contribution

It proposes a novel superpixel generation and clustering pipeline that enhances weakly supervised tumor segmentation using only image-level labels.

Findings

Achieved a mean Dice coefficient of 0.691 on the test set.

Outperformed existing superpixel-based weakly supervised methods.

Demonstrated effectiveness on brain MRI data from the BraTS 2020 dataset.

Abstract

Training machine learning models to segment tumors and other anomalies in medical images is an important step for developing diagnostic tools but generally requires manually annotated ground truth segmentations, which necessitates significant time and resources. This work proposes the use of a superpixel generation model and a superpixel clustering model to enable weakly supervised brain tumor segmentations. The proposed method utilizes binary image-level classification labels, which are readily accessible, to significantly improve the initial region of interest segmentations generated by standard weakly supervised methods without requiring ground truth annotations. We used 2D slices of magnetic resonance brain scans from the Multimodal Brain Tumor Segmentation Challenge 2020 dataset and labels indicating the presence of tumors to train the pipeline. On the test cohort, our method achieved a mean Dice coefficient of 0.691 and a mean 95% Hausdorff distance of 18.1, outperforming existing superpixel-based weakly supervised segmentation methods.