arXiv cs.CV
· Papers
MRIComp4Flow: Compression of 3D Brain MRI for Training Multi-Modal Generative Models
arXiv:2608.10291v1 Announce Type: new Abstract: Large-scale multi-modal MRI datasets impose substantial storage and I/O costs, limiting the training of 3D generative models on commodity infrastructure. While lossy compression is known to preserve accuracy for discriminative segmentation networks, its effect on generati