Kwark et al. COMB framework for label-free whole-slide virtual staining at UIUC, USA; Significant gains in fidelity and tiling consistency.
Kwark et al. COMB framework for label-free whole-slide virtual staining at UIUC, USA; Significant gains in fidelity and tiling consistency. Seamless Whole Slide Label-Free Virtual Staining Seamless Whole Slide Label-Free Virtual Staining Dou Hoon Kwark1, Kianoush Falahkheirkhah1, Ji-Hun Oh1, Shirui Luo2, Volodymyr Kindratenko1,2⋆, and Rohit Bhargava1⋆ 1 University of Illinois Urbana-Champaign, Urbana, IL, USA 2 National Center for Supercomputing Applications, Urbana, IL, USA dkwark2@illinois.edu Abstract. Label-free virtual staining offers a compelling, non-destruct- ive alternative to standard histopathology; however, its clinical adoption is hindered by the computational bottlenecks inherent to processing gi- gapixel Whole Slide Images (WSIs). Current deep learning approaches re- quire patch-based inference to avoid memory constraints, which disrupts global tissue continuity and introduces tiling artifacts—displaying visi- ble seams and color shifts. To address this, we introduce the Consistency Memory Bank (COMB), a novel label-free virtual staining framework that enforces spatial and channel consistency across tiles without memory bottlenecks. COMB decouples context storage from computation, utiliz- ing a dynamic retrieval mechanism to fetch feature representations from adjacent tiles. This enables a retrieval-based context integration strategy that adopts local padding to resolve spatial discontinuities and neighbor- aware channel attention to stabilize statistical drift. Further optimized with a sliding window schedule to ensure minimal memory overhead, our method demonstrates superior performance over state-of-the-art base- lines, achieving significant improvements in both perceptual fidelity and tiling consistency, while suggesting its downstream utility in tumor seg- mentation. Code is available at https://github.com/dou0000/COMB. Keywords: Label-free virtual staining · Tiling artifact mitigation 1 Introduction Pathology diagnosis relies on physical staining to make tissue structures visible under a microscope. While effective, this standard workflow is labor-intensive, inconsistent due to chemical variations, and consumes tissue samples required for downstream molecular assays. Label-free imaging [6] offers a compelling al- ternative by visualizing tissue structures using their inherent signals, eliminating the need for destructive chemical dyeing process. The advent of virtual stain- ing—using deep learning to translate raw data into histological images—offers a powerful way to optimize pathology workflows across diverse imaging modalities [22,24,7,8,9] However, clinical adoption faces a major computational hurdle: memory con- straints prevent the single holistic inference of gigapixel Whole-Slide Images ⋆Contributed equally arXiv:2609.10914v1 [eess.IV] 9 Sep 2026 ...