Abstract
Large-scale multi-modal MRI datasets impose substantial storage and I/O costs, limiting the training of 3D generative models on commodityinfrastructure.Whilelossycompressionisknowntopreserve accuracy for discriminative segmentation networks, its effect on generative models, which must learn the full data distribution rather than a decision boundary, is unexplored. We study whether standard image codecscaneffectivelycompresssemanticallyrichbraintumorMRIwhile preservingthefidelityrequiredtotrainanddeploya3DMRIgenerative model.Each3DvolumeiscompressedwithJPEG2000oranear-lossless JPEG-LS pipeline. Next, a Wavelet Flow Matching model, conditioned onBraTSimagesequences(T1n,T1c,T2,T2f),istrainedoncompressed data, and the resulting models are evaluated on the validation set. At a 20:1 compression ratio, synthesis quality is statistically equivalent to a model trained on uncompressed data within a pre-specified margin (ΔPSNR<1dB,ΔSSIM<0.02;pairedTOSTp=[[p]]):meanPSNRis 27.3dBvs.27.0dBandmeanSSIMis0.95vs.0.96acrossmodalities.Our results indicate that JPEG2000 compression is a practical step toward scalable3DMRIgenerativemodelingwithoutdegradingsynthesisquality.Thecodeisavailableathttps://github.com/lisafis/MRIComp4Flow.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/SASHIMI_055.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=RJ1TBjrsZF
BibTex
@InProceedings{FisLis_MRIComp4Flow_MICCAISAT2026,
author = { Fischer, Lisa K. AND Riabets, Mykhailo AND Rueckert, Daniel AND Wiestler, Benedikt AND Meyer-Baese, Anke AND Nagar, Sandeep},
title = { { MRIComp4Flow: Compression of 3D Brain MRI for Training Multi-Modal Generative Models } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17258},
month = {pending},
page = {pending}
}