German Cancer Research Center (DKFZ)

Open-weights pathology foundation models pairing a 1.1B-parameter ViT-g/8 tile encoder with distilled 86M and 22M variants for H&E histology.

Histopathology model predicting 102 methylation-defined CNS tumor subtypes from H&E whole-slide images, with calibrated per-case confidence scores.

Vision-language encoders for chest CT that align 3D volumes with radiology reports using contrastive, report-generation, and masked-image objectives.

Phylogeny-aware genomic language model scoring variant effects from whole-genome alignments and species trees across three evolutionary timescales.

Complex-valued diffusion model generating synthetic MRI k-space phase from magnitude images, raising k-space skull-stripping Dice from 41.1% to 80.1%.

Whole-slide pathology embedding framework that ranks tiles, keeps only the 25 most informative, and encodes a slide in 2.27 seconds.

Masked-autoencoder foundation model that pre-trains a 3D Residual Encoder U-Net on roughly 39,000 brain MRIs for volumetric image segmentation.

Self-supervised pathology foundation model with a 300M-parameter vision transformer tile encoder, trained on a multi-stain whole-slide image corpus.

Cryo-electron tomography membrane analysis pipeline pairing generalizable U-Net membrane segmentation with mesh-based particle localization.