arXiv subjects

Songhao Li

Publications and source records attributed to Songhao Li.

8 recordsLinked to original sources

arXiv2026

CytoFormer: A Molecularly Supervised Cell Foundation Model for Histopathology Cell Classification

Identifying cell types directly from routine haematoxylin and eosin (H&E) histology would enable single-cell analysis at scale, but training such models has relied on manual pathologist annotations, which are slow, expensive and unreliable for many cell types. We instead supervise morphology with molecules. Imaging-based spatial transcriptomics profiles individual cells in situ on a section that can afterwards be stained with H&E, so that molecular identity and morphology are observed for the same physical cell. We assembled 81 such paired Xenium sections spanning 16 organs, derived per-cell labels by clustering, marker-gene annotation, organ-wise human review and quality control, and mapped them onto the cell types commonly reported in each organ. This yielded 15.4 million cells, each with a paired H&E image patch and one of 23 cell types, on which we trained CytoFormer, a cell foundation model with a multi-task, per-organ classification head. On spatially held-out tissue CytoFormer reached an accuracy of 0.85 and a macro-F1 of 0.78 across all 16 organs, and its predictions reproduced the tissue architecture of an entire held-out section. The representation also transfers: with the encoder frozen, a linear head on CytoFormer features performed better than six pathology foundation models on four expert-annotated benchmarks, including on organs and cell types that were not part of pretraining. Finally, in an interactive active-learning setting, CytoFormer's embeddings are markedly more label-efficient than existing pathology foundation models, detecting normal epithelium amid look-alike tumour with an F1 of 0.82 from only a few annotations and leading the strongest baseline by 0.13 in F1. CytoFormer turns paired H&E and spatial transcriptomics into a reusable, label-efficient representation for cell-level analysis of routine histology.

cs.CV↗

arXiv2026

VISTA-PATH: An interactive foundation model for pathology image segmentation and quantitative analysis in computational pathology

Accurate semantic segmentation for histopathology image is crucial for quantitative tissue analysis and downstream clinical modeling. Recent segmentation foundation models have improved generalization through large-scale pretraining, yet remain poorly aligned with pathology because they treat segmentation as a static visual prediction task. Here we present VISTA-PATH, an interactive, class-aware pathology segmentation foundation model designed to resolve heterogeneous structures, incorporate expert feedback, and produce pixel-level segmentation that are directly meaningful for clinical interpretation. VISTA-PATH jointly conditions segmentation on visual context, semantic tissue descriptions, and optional expert-provided spatial prompts, enabling precise multi-class segmentation across heterogeneous pathology images. To support this paradigm, we curate VISTA-PATH Data, a large-scale pathology segmentation corpus comprising over 1.6 million image-mask-text triplets spanning 9 organs and 93 tissue classes. Across extensive held-out and external benchmarks, VISTA-PATH consistently outperforms existing segmentation foundation models. Importantly, VISTA-PATH supports dynamic human-in-the-loop refinement by propagating sparse, patch-level bounding-box annotation feedback into whole-slide segmentation. Finally, we show that the high-fidelity, class-aware segmentation produced by VISTA-PATH is a preferred model for computational pathology. It improve tissue microenvironment analysis through proposed Tumor Interaction Score (TIS), which exhibits strong and significant associations with patient survival. Together, these results establish VISTA-PATH as a foundation model that elevates pathology image segmentation from a static prediction to an interactive and clinically grounded representation for digital pathology. Source code and demo can be found at https://github.com/zhihuanglab/VISTA-PATH.

cs.CV↗

arXiv2025

A co-evolving agentic AI system for medical imaging analysis

Agentic AI is rapidly advancing in healthcare and biomedical research. However, in medical image analysis, their performance and adoption remain limited due to the lack of a robust ecosystem, insufficient toolsets, and the absence of real-time interactive expert feedback. Here we present "TissueLab", a co-evolving agentic AI system that allows researchers to ask direct questions, automatically plan and generate explainable workflows, and conduct real-time analyses where experts can visualize intermediate results and refine them. TissueLab integrates tool factories across pathology, radiology, and spatial omics domains. By standardizing inputs, outputs, and capabilities of diverse tools, the system determines when and how to invoke them to address research and clinical questions. Across diverse tasks with clinically meaningful quantifications that inform staging, prognosis, and treatment planning, TissueLab achieves state-of-the-art performance compared with end-to-end vision-language models (VLMs) and other agentic AI systems such as GPT-5. Moreover, TissueLab continuously learns from clinicians, evolving toward improved classifiers and more effective decision strategies. With active learning, it delivers accurate results in unseen disease contexts within minutes, without requiring massive datasets or prolonged retraining. Released as a sustainable open-source ecosystem, TissueLab aims to accelerate computational research and translational adoption in medical imaging while establishing a foundation for the next generation of medical AI.

cs.CV↗

arXiv2018

Symplectic groupoids for cluster manifolds

We construct symplectic groupoids integrating log-canonical Poisson structures on cluster varieties of type $\mathcal{A}$ and $\mathcal{X}$ over both the real and complex numbers. Extensions of these groupoids to the completions of the cluster varieties where cluster variables are allowed to vanish are also considered. In the real case, we construct source-simply-connected groupoids for the cluster charts via the Poisson spray technique of Crainic and M\u{a}rcu\c{t}. These groupoid charts and their analogues for the symplectic double and blow-up groupoids are glued by lifting the cluster mutations to groupoid comorphisms whose formulas are motivated by the Hamiltonian perspective of cluster mutations introduced by Fock and Goncharov.

math.SG↗

arXiv2015

Lagrangian Mechanics and Reduction on Fibered Manifolds

This paper develops a generalized formulation of Lagrangian mechanics on fibered manifolds, together with a reduction theory for symmetries corresponding to Lie groupoid actions. As special cases, this theory includes not only Lagrangian reduction (including reduction by stages) for Lie group actions, but also classical Routh reduction, which we show is naturally posed in this fibered setting. Along the way, we also develop some new results for Lagrangian mechanics on Lie algebroids, most notably a new, coordinate-free formulation of the equations of motion. Finally, we extend the foregoing to include fibered and Lie algebroid generalizations of the Hamilton-Pontryagin principle of Yoshimura and Marsden, along with the associated reduction theory.

math.DS↗

arXiv2014

The tropical momentum map: a classification of toric log symplectic manifolds

We give a generalization of toric symplectic geometry to Poisson manifolds which are symplectic away from a collection of hypersurfaces forming a normal crossing configuration. We introduce the tropical momentum map, which takes values in a generalization of affine space called a log affine manifold. Using this momentum map, we obtain a complete classification of such manifolds in terms of decorated log affine polytopes, hence extending the classification of symplectic toric manifolds achieved by Atiyah, Guillemin-Sternberg, Kostant, and Delzant.

math.SG↗

arXiv2013

The Stokes Groupoids

We construct and describe a family of groupoids over complex curves which serve as the universal domains of definition for solutions to linear ordinary differential equations with singularities. As a consequence, we obtain a direct, functorial method for resumming formal solutions to such equations.

math.AG↗

arXiv2012

Symplectic groupoids of log symplectic manifolds

A log symplectic manifold is a Poisson manifold which is generically nondegenerate. We develop two methods for constructing the symplectic groupoids of log symplectic manifolds. The first is a blow-up construction, corresponding to the notion of an elementary modification of a Lie algebroid along a subalgebroid. The second is a gluing construction, whereby groupoids defined on the open sets of an appropriate cover may be combined to obtain global integrations. This allows us to classify all Hausdorff symplectic groupoids of log symplectic manifolds in a combinatorial fashion, in terms of a certain graph of fundamental groups associated to the manifold. Using the same ideas, and as a first step, we also construct and classify the groupoids integrating the Lie algebroid of vector fields tangent to a smooth hypersurface.

math.SG↗