Synthetic Generation and Latent Projection Denoising of Rim Lesions in Multiple Sclerosis
Not on the map. Why?
Abstract (arXiv)
Quantitative susceptibility maps from magnetic resonance images can provide both prognostic and diagnostic information in multiple sclerosis, a neurodegenerative disease characterized by the formation of lesions in white matter brain tissue. In particular, susceptibility maps provide adequate contrast to distinguish between "rim" lesions, surrounded by deposited paramagnetic iron, and "non-rim" lesion types. These paramagnetic rim lesions (PRLs) are an emerging biomarker in multiple sclerosis. Much effort has been devoted to both detection and segmentation of such lesions to monitor longitudinal change. As paramagnetic rim lesions are rare, addressing this problem requires confronting the class imbalance between rim and non-rim lesions. We produce synthetic quantitative susceptibility maps of paramagnetic rim lesions and show that inclusion of such synthetic data improves classifier performance and provide a multi-channel extension to generate accompanying contrasts and probabilistic segmentation maps. We exploit the projection capability of our trained generative network to demonstrate a novel denoising approach that allows us to train on ambiguous rim cases and substantially increase the minority class. We show that both synthetic lesion synthesis and our proposed rim lesion label denoising method best approximate the unseen rim lesion distribution and improve detection in a clinically interpretable manner. We release our code and generated data at https://github.com/agr78/PRLx-GAN upon publication.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 21 of its 61 references that are in the map. First posted on arXiv 29 May 2025.
Builds on
21 of 61 references.
References · 21
- 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
- 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
- 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
- 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,588
- 2015Learning Deep Features for Discriminative Localizationcited by 2,436
- 2014Conditional Generative Adversarial Netscited by 2,162
- 2020Training Generative Adversarial Networks with Limited Datacited by 1,186
- 2016Inverting The Generator Of A Generative Adversarial Networkcited by 140
- 2021CovidGAN: Data Augmentation Using Auxiliary Classifier GAN for Improved Covid-19 Detectioncited by 64
- 2020Do 2D GANs Know 3D Shape? Unsupervised 3D shape reconstruction from 2D Image GANscited by 63
- 2021SinGAN-Seg: Synthetic training data generation for medical image segmentationcited by 43
- 2018Robustness of Conditional GANs to Noisy Labelscited by 27
9 more
- 2021Transfer Learning in Magnetic Resonance Brain Imaging: a Systematic Reviewcited by 19
- 2021Prior Image-Constrained Reconstruction using Style-Based Generative Modelscited by 19
- 2019Multiple Sclerosis Lesion Synthesis in MRI using an encoder-decoder U-NETcited by 13
- 2021Improving Medical Image Classification with Label Noise Using Dual-uncertainty Estimationcited by 13
- 2023SYNAuG: Exploiting Synthetic Data for Data Imbalance Problemscited by 13
- 2021A Mixed-Supervision Multilevel GAN Framework for Image Quality Enhancementcited by 11
- 2023Maximum Spherical Mean Value (mSMV) Filtering for Whole Brain Quantitative Susceptibility Mappingcited by 2
- 2024QSM-RimDS: A detection and segmentation tool for paramagnetic rim lesions in multiple sclerosiscited by 1
- 2025Correcting Noisy Multilabel Predictions: Modeling Label Noise through Latent Space Shiftscited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.