BCD-Net Model Card

BCD-Net is a Deep Variational Bayesian Neural Network for the Blind Color Deconvolution (BCD) of histological images. It allows to separate a histological image of any size into the color matrix and the concentration matrices. The weights are released in the format of a PyTorch state dictionary.

Original Github repository

The code and instructions to build the network and make predictions are available at Github.

Our paper: Deep Bayesian Blind Color Deconvolution of Histological Images

For more information about the BCD-Net methodology and training procedure, please have a look at our paper.

Citation

If you find our model helpful, please consider citing our paper:

@article{,
    title={{D}eep {B}ayesian {B}lind {C}olor {D}econvolution of {H}istological {I}mages},
    author={Shuowen Yang and Fernando Pérez-Bueno and Francisco M. Castro-Macías and Rafael Molina and Aggelos K. Katsaggelos},
    archivePrefix={arXiv},
    primaryClass={},
    year={2023}
}