Beer-Lambert Autoencoder for Unsupervised Stain Representation Learning and Deconvolution in Multi-immunohistochemical Brightfield Histology Images

TL;DR

This paper introduces a data-driven, unsupervised autoencoder model that improves stain separation in multiplex immunohistochemistry images, enabling better deconvolution and analysis of multiple chromogens beyond traditional methods.

Contribution

It presents a novel U-Net based encoder with a differentiable Beer-Lambert model decoder for unsupervised stain deconvolution in multiplex IHC images, outperforming classical matrix-based approaches.

Findings

Achieves high-quality RGB reconstruction of multiplex IHC images.

Significantly reduces inter-channel bleed-through compared to traditional methods.

Demonstrates effective separation of five chromogens in colorectal tissue samples.

Abstract

Separating the contributions of individual chromogenic stains in RGB histology whole slide images (WSIs) is essential for stain normalization, quantitative assessment of marker expression, and cell-level readouts in immunohistochemistry (IHC). Classical Beer-Lambert (BL) color deconvolution is well-established for two- or three-stain settings, but becomes under-determined and unstable for multiplex IHC (mIHC) with K>3 chromogens. We present a simple, data-driven encoder-decoder architecture that learns cohort-specific stain characteristics for mIHC RGB WSIs and yields crisp, well-separated per-stain concentration maps. The encoder is a compact U-Net that predicts K nonnegative concentration channels; the decoder is a differentiable BL forward model with a learnable stain matrix initialized from typical chromogen hues. Training is unsupervised with a perceptual reconstruction objective augmented by loss terms that discourage unnecessary stain mixing. On a colorectal mIHC panel comprising 5 stains (H, CDX2, MUC2, MUC5, CD8) we show excellent RGB reconstruction, and significantly reduced inter-channel bleed-through compared with matrix-based deconvolution. Code and model are available at https://github.com/measty/StainQuant.git.