colour deconvolution

colour deconvolution decomposes RGB microscopy images into dye-specific optical absorbance and transmittance channels to enable quantitative analysis of stained histological samples.

Key Features:

• Stain unmixing algorithm: Implements stain unmixing based on Ruifrok and Johnston’s method to separate individual dye contributions.
• RGB decomposition: Converts RGB images into separate channels that represent optical absorbance and transmittance properties of specific dyes.
• Numerical colour and intensity processing: Performs numerical analysis of colour distribution and intensity for quantitative extraction of stain information.
• Downstream analysis outputs: Produces dye-specific channels suitable for morphological and histochemical segmentation, automated marker localization, and image enhancement.
• Implementations: Provided as an ImageJ plugin written in Java and as a MATLAB program/function.

Scientific Applications:

• Morphological segmentation: Enables segmentation of tissue structures using dye-specific absorbance channels.
• Histochemical segmentation: Supports separation and quantitative assessment of histochemical stains.
• Automated marker localization: Facilitates localization of immunohistochemical or other markers by isolating stain signals.
• Image enhancement and quantitation: Improves contrast and enables quantitative measurement of staining intensities in digital histopathology images.

Methodology:

Applies stain unmixing algorithms based on Ruifrok and Johnston’s method to convert RGB images into optical absorbance/transmittance channels and performs numerical processing of colour distribution and intensity; implemented as an ImageJ plugin (Java) and a MATLAB program/function.