Image-to-image reconstruction problems with free or inexpensive metadata in the form of class labels appear often in biological and medical image domains. Existing text-guided or style-transfer image-to-image approaches do not translate to datasets where additional information is provided as discrete classes. We introduce and implement a model which combines image-to-image and class-guided denoising diffusion probabilistic models. We train our model on a real-world dataset of microscopy images used for drug discovery, with and without incorporating metadata labels. By exploring the properties of image-to-image diffusion with relevant labels, we show that class-guided image-to-image diffusion can improve the meaningful content of the reconstructed images and outperform the unguided model in useful downstream tasks.
@article{arxiv.2303.08863,
title = {Class-Guided Image-to-Image Diffusion: Cell Painting from Brightfield Images with Class Labels},
author = {Jan Oscar Cross-Zamirski and Praveen Anand and Guy Williams and Elizabeth Mouchet and Yinhai Wang and Carola-Bibiane Schönlieb},
journal= {arXiv preprint arXiv:2303.08863},
year = {2023}
}