Method for analyzing immunohistochemistry images
AssigneeNational Cheng Kung University
InventorsPau-Choo Chung, Hung-Wen Tsai, Shao-Kai Liao, Kuo-Sheng Cheng
Life of this patent
Estimating the term…
Abstract
A method for analyzing an immunohistochemistry (IHC) image is provided and includes: segmenting nuclei from the IHC image according to a machine learning model; removing pixels belonging to the nuclei and pixels in a color range from the IHC image to obtain multiple cytoplasmic pixels; assign the cytoplasmic pixels to the nuclei to form multiple cells according to the locations of the cytoplasmic pixels; and calculate a pixel staining score of each pixel in the cells, thereby calculating a cell staining score for each cell.
Claims15
A method performed by a computer system for analyzing an immunohistochemistry image, the method comprising:
- segmenting a plurality of cell nuclei from the immunohistochemistry image according to a machine learning wherein the model, immunohistochemistry image comprises a plurality of pixels, and each of the pixels comprises a plurality of color channels;
- removing the pixels belonging to the cell nuclei while outlines of the cell nuclei are retained, transforming each of the pixels of the immunohistochemistry image into a color space comprising a phase, saturation and brightness, and removing the pixels from the immunohistochemistry image in at least one color range to obtain a plurality of cytoplasmic pixels accordingly, wherein the cytoplasmic pixels are free of the pixels belonging to the cell nuclei and free of the pixels of the immunohistochemistry image in the at least one color range;
- assigning each of the cytoplasmic pixels to one of the cell nuclei according to a location of the corresponding cytoplasmic pixel to form a plurality of cells;
- and
- calculating a pixel staining score of each of the pixels in each of the cells to calculate a cell staining score of each of the cells.
An electrical device comprising:
- a memory configured to store a plurality of instructions;
- and
- a processor configured to execute the instructions to perform a plurality of steps:
- segmenting a plurality of cell nuclei from a immunohistochemistry image according to a machine learning model, wherein the immunohistochemistry image comprises a plurality of pixels, and each of the pixels comprises a plurality of color channels;
- removing the pixels belonging to the cell nuclei while outlines of the cell nuclei are retained, transforming each of the pixels of the immunohistochemistry image into a color space comprising a phase, saturation and brightness, and removing the pixels from the immunohistochemistry image in at least one color range to obtain a plurality of cytoplasmic pixels accordingly, wherein the cytoplasmic pixels are free of the pixels belonging to the cell nuclei and free of the pixels of the immunohistochemistry image in the at least one color range;
- assigning each of the cytoplasmic pixels to one of the cell nuclei according to a location of the corresponding cytoplasmic pixel to form a plurality of cells;
- and
- calculating a pixel staining score of each of the pixels in each of the cells to calculate a cell staining score of each of the cells.
A non-transitory computer readable storage media storing a plurality of instructions which are configured to be executed by a computer system to perform a plurality of steps:
- segmenting a plurality of cell nuclei from a immunohistochemistry image according to a machine learning model, wherein the immunohistochemistry image comprises a plurality of pixels, and each of the pixels comprises a plurality of color channel;
- removing the pixels belonging to the cell nuclei while outline of the cell nuclei are retained, transforming each of the pixels of the immunohistochemistry image into a color space comprising a phase, saturation and brightness, and removing the pixels from the immunohistochemistry image in at least one color range to obtain a plurality of cytoplasmic pixels accordingly, wherein the cytoplasmic pixels are free of the pixels belonging to the cell nuclei and free of the pixels of the immunohistochemistry image in the at least one color range;
- assigning each of the cytoplasmic pixels to one of the cell nuclei according to a location of the corresponding cytoplasmic pixel to form a plurality of cells;
- and
- calculating a pixel staining score of each of the pixels in each of the cells to calculate a cell staining score of each of the cells.
Dependencies are inferred from the claim text.