[Submitted on 5 Jun 2026 (v1), last revised 14 Oct 2026 (this version, v3)]
Abstract:Synthetic histopathology image generation addresses patient-privacy concerns and the growing data demands of foundation models. Existing state-of-the-art histopathology generative models use pretrained Vision Foundation Models (VFMs) as conditioning signals. We show this yields conditioning-dominated diversity: on TCGA-BRCA, 62-75% of their output diversity is attributable to the conditioning signal rather than the learned latent space, while de novo synthesis still requires a VFM at inference. We instead use histopathology VFMs as the latent space itself: their patch tokens are $ell_2$-normalized on the unit hypersphere $mathcal{S}^{d-1}$ with strong angular dominance and intrinsic curvature, motivating a Riemannian formulation. We present STREAM, the first framework to apply Riemannian flow matching in the histopathology domain, in two stages: 1) a bridge-type stochastic perturbation that establishes per-token rectifiability on $mathcal{S}^{d-1}$ for training a Diffusion Transformer, and 2) a novel decoder training design whose noise covariance is anisotropic in the left-singular basis of the per-token tangent-projected velocity-field Jacobian, spending a large robustness budget on its low-response directions and a small one on its high-response directions. Across TCGA-BRCA and TCGA-COADREAD, STREAM achieves state-of-the-art gFID and ranks first on nearly all histopathology-specific metrics as well. Code and a public gallery of generated images are available at this https URL.
Submission history
From: Won June Cho [view email]
[v1] Fri, 5 Jun 2026 08:26:44 UTC (8,031 KB)
[v2] Mon, 5 Oct 2026 12:17:47 UTC (18,543 KB)
[v3] Wed, 14 Oct 2026 05:30:57 UTC (18,543 KB)