π€ AI Summary
The core challenge in virtual staining lies in tissue deformation induced by chemical staining, which impedes acquisition of pixel-wise aligned paired training dataβleading existing methods to rely heavily on strong supervision and suffer from poor generalizability. This paper proposes a misalignment-robust generative virtual staining framework that innovatively integrates generative adversarial networks with learnable cascaded spatial transformer networks. It achieves end-to-end joint optimization of registration and staining through multi-stage deformation correction, significantly reducing dependence on strictly paired data. The method maintains high accuracy even under unpaired or weakly paired conditions. Evaluated across five datasets, it consistently outperforms state-of-the-art approaches: average PSNR improves by 3.2% on internal and 10.1% on external test sets; under severe misalignment, PSNR gain reaches 23.8%, markedly enhancing clinical applicability.
π Abstract
Accurate histopathological diagnosis often requires multiple differently stained tissue sections, a process that is time-consuming, labor-intensive, and environmentally taxing due to the use of multiple chemical stains. Recently, virtual staining has emerged as a promising alternative that is faster, tissue-conserving, and environmentally friendly. However, existing virtual staining methods face significant challenges in clinical applications, primarily due to their reliance on well-aligned paired data. Obtaining such data is inherently difficult because chemical staining processes can distort tissue structures, and a single tissue section cannot undergo multiple staining procedures without damage or loss of information. As a result, most available virtual staining datasets are either unpaired or roughly paired, making it difficult for existing methods to achieve accurate pixel-level supervision. To address this challenge, we propose a robust virtual staining framework featuring cascaded registration mechanisms to resolve spatial mismatches between generated outputs and their corresponding ground truth. Experimental results demonstrate that our method significantly outperforms state-of-the-art models across five datasets, achieving an average improvement of 3.2% on internal datasets and 10.1% on external datasets. Moreover, in datasets with substantial misalignment, our approach achieves a remarkable 23.8% improvement in peak signal-to-noise ratio compared to baseline models. The exceptional robustness of the proposed method across diverse datasets simplifies the data acquisition process for virtual staining and offers new insights for advancing its development.