AAAI 2026 Accepted

This repository contains the official implementation of the paper:

Virtual Multiplex Staining for Histological Images using a Marker-wise Conditioned Diffusion Model
Hyun-Jic Oh, Junsik Kim, Zhiyi Shi, Yichen Wu, Yu-An Chen, Peter K. Sorger, Hanspeter Pfister, Won-Ki Jeong

We propose a marker-wise conditioned latent diffusion framework that generates virtual multiplex (mIF/mIHC) marker channels directly from corresponding H&E images while sharing a single unified architecture across all markers. The model supports marker-by-marker synthesis, accommodates heterogeneous marker intensity distributions, and is fine-tuned for single-step inference to improve both visual fidelity and runtime efficiency.

If you find this work useful in your research, please consider citing our paper:

@article{oh2025virtual,
  title   = {Virtual Multiplex Staining for Histological Images using a Marker-wise Conditioned Diffusion Model},
  author  = {Oh, Hyun-Jic and Kim, Junsik and Shi, Zhiyi and Wu, Yichen and Chen, Yu-An and Sorger, Peter K and Pfister, Hanspeter and Jeong, Won-Ki},
  journal = {arXiv preprint arXiv:2508.14681},
  year    = {2025}
}

• data loader
• training
• inference
• preprocessing/postprocessing parts

Our implementation, training scripts, and evaluation pipelines heavily draw inspiration from Marigold and diffusion-e2e-ft, and we gratefully acknowledge their authors for releasing high-quality code and models.