Self-supervised Brain Lesion Generation for Effective Data Augmentation of Medical Images

TL;DR

This paper introduces a novel self-supervised framework for generating synthetic brain lesions to augment training data, significantly improving segmentation accuracy in medical imaging.

Contribution

It presents a new lesion generator using adversarial autoencoders, a Soft Poisson Blending algorithm for seamless image composition, and a prototype consistency regularization for better model training.

Findings

Improved Dice score from 50.36% to 60.23% on ATLAS v2.0 dataset.

Outperforms existing data augmentation methods.

Validated on two public datasets with consistent improvements.

Abstract

Accurate brain lesion delineation is important for planning neurosurgical treatment. Automatic brain lesion segmentation methods based on convolutional neural networks have demonstrated remarkable performance. However, neural network performance is constrained by the lack of large-scale well-annotated training datasets. In this manuscript, we propose a comprehensive framework to efficiently generate new samples for training a brain lesion segmentation model. We first train a lesion generator, based on an adversarial autoencoder, in a self-supervised manner. Next, we utilize a novel image composition algorithm, Soft Poisson Blending, to seamlessly combine synthetic lesions and brain images to obtain training samples. Finally, to effectively train the brain lesion segmentation model with augmented images we introduce a new prototype consistence regularization to align real and synthetic features. Our framework is validated by extensive experiments on two public brain lesion segmentation datasets: ATLAS v2.0 and Shift MS. Our method outperforms existing brain image data augmentation schemes. For instance, our method improves the Dice from 50.36% to 60.23% compared to the U-Net with conventional data augmentation techniques for the ATLAS v2.0 dataset.