OTLesMix: Wasserstein Barycenter and Optimal Transport Map for Synthetic Lesion Generation with Diverse Shapes and Locations

2026-08-06

AI Summary

Existing medical image synthesis methods struggle to simultaneously preserve the diversity of lesion shape and spatial location, limiting the effectiveness of data augmentation in improving segmentation model performance. This work proposes OTLesMix, a novel approach that, for the first time, integrates Wasserstein barycenters with optimal transport mappings to generate synthetic lesions exhibiting diverse morphologies and spatial distributions derived from real pathological instances. The method employs a mixing strategy for lesion synthesis and incorporates these augmented samples into the training of deep segmentation models. Evaluated on three brain lesion segmentation tasks, OTLesMix achieves Dice score improvements of 2.9–6.6 points over non-augmented baselines and significantly outperforms state-of-the-art mix-based augmentation techniques.

Abstract

The development of deep learning over the past decade has revolutionized medical imaging segmentation, allowing the extraction of precise descriptors from large volumes to characterize pathologies. Data augmentation is a technique widely regarded as a way to improve model training. It includes simple transformations like spatial operations or intensity modifications, but also more advanced synthesis techniques. Their goal is to generate new realistic samples from an existing dataset to diversify the images used during training. Among them, several propose different mixing strategies to combine real samples. However, one of their major shortcomings is to yield limited variability in terms of generated lesion shapes and locations. In this work, we introduce a novel image synthesis method, called OTLesMix, that leverages Wasserstein barycenter and optimal transport plan to generate realistic and diverse samples. We evaluated our method on three brain lesion segmentation tasks, on which it improves the Dice score compared to a model trained without synthetic data by 2.9 to 6.6 points, and outperforms state-of-the-art mix-based methods.

Problem

• synthetic lesion generation
• lesion shape diversity
• lesion location variability
• data augmentation
• medical image synthesis

Innovation

• Wasserstein barycenter
• optimal transport
• synthetic lesion generation
• data augmentation
• medical image segmentation