Back to results
US 11,107,248 B2Granted
Color unmixing with scatter correction
AssigneeVentana Medical Systems, Inc.
InventorYao Nie
Life of this patent
Estimating the term…
Abstract
The present disclosure provides systems and methods for separating colors in an image by automatically selecting color reference vectors that take into consideration the effect of light scattering, and principally how the light scattering changes the proportions of RGB channel signals in detected light at varying stain concentrations.
Claims20
A method of unmixing an image of a biological specimen stained with at least two stains comprising:
- deriving total optical density values for each channel in R, G, B color space from pixel intensity values in an image or a portion of an image having a first stain and a second stain, wherein at least the first stain is a concentration-dependent stain;
- obtaining a plurality of prospective first stain color reference vectors, each of the plurality of prospective first stain color reference vectors characterizing the first stain at a different concentration;
- obtaining at least one second stain color reference vector;
- deriving a series of candidate color systems from the first stain color and second stain color reference vectors, each candidate color system comprising one of the prospective first stain color reference vectors from the plurality of prospective first stain color reference vectors and the at least one second stain color reference vector;
- selecting a color system from the series of candidate color system, the color system selected by:
- computing a reconstruction error for each candidate color space;
- and
- determining the candidate color space having a minimal reconstruction error;
- and
- unmixing the acquired image using the selected color system.
7 more elements
A system for unmixing an image of a biological specimen stained with at least two stains, the system comprising:
- one or more processors;
- and
- one or more memories coupled to the one or more processors, the one or more memories to store computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
- acquiring an image of the biological sample having signals corresponding to a first stain and a second stain, wherein at least the first stain is a concentration-dependent stain;
- deriving total optical density values for each channel in R, G, B color space from pixel intensity values in the acquired image;
- obtaining a set of prospective first stain color reference vectors from a spectral reference database, each color reference vector within the set of prospective first stain color reference vectors characterizing the first stain at a different stain concentration;
- obtaining at least one second stain color reference vector from the spectral reference database;
- deriving a series of candidate color systems from the first stain and second stain color reference vectors, each candidate color system comprising one of the prospective first stain color reference vectors from the set of prospective first stain color reference vectors and the at least one second stain color reference vector;
- selecting a color system from the series of candidate color system, the color system selected by:
- computing a reconstruction error for each candidate color space;
- and
- determining the candidate color space having a minimal reconstruction error;
- and
- unmixing the signals within the acquired image using the selected color system.
11 more elements
A non-transitory computer-readable medium storing instructions which, when executed by one or more processors of an imaging system, cause the imaging system to:
- derive total optical density values for each channel in R, G, B color space from pixel intensity values in an image or a portion of an image having a first stain and a second stain, wherein at least the first stain is a concentration-dependent stain;
- obtain a plurality of prospective first stain color reference vectors, each of the plurality of prospective first stain color reference vectors characterizing the first stain at a different concentration;
- obtain at least one second stain color reference vector;
- derive a series of candidate color systems from the first stain color and second stain color reference vectors, each candidate color system comprising one of the prospective first stain color reference vectors from the plurality of prospective first stain color reference vectors and the at least one second stain color reference vector;
- select a color system from the series of candidate color system, the color system selected by:
- computing a reconstruction error for each candidate color space;
- and
- determining the candidate color space having a minimal reconstruction error;
- and
- unmix the acquired image using the selected color system.
7 more elements
Dependencies are inferred from the claim text.