ROOF: Robustness-Oriented Optimal Factorization for Fluorescence Crosstalk Correction.

ROOF, a semi-blind diagonally normalized nonnegative matrix factorization method, achieves higher crosstalk correction accuracy on four-band fluorescence images while suppressing overcorrection artifacts.

• Why it matters: Fluorescence bleed-through in multi-band microscopy can distort quantification and visualization, and existing calibration or blind unmixing approaches can be brittle, require manual tuning, or introduce overcorrection.
• What they did: ROOF uses a reference crosstalk matrix from single-labeled controls and estimates the regularization parameter automatically via maximum likelihood, avoiding manual parameter selection.
• The result: On synthetic and real four-band fluorescence images, ROOF consistently outperforms state-of-the-art methods, improving accuracy, preserving biologically meaningful colocalization, and remaining computationally efficient.

AI summary of the abstract · may contain errors · not medical advice · read the paper.