Pathological diagnosis support

• US 8,787,650 B2
• Filed: 01/23/2009
• Issued: 07/22/2014
• Est. Priority Date: 01/29/2008
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Marugame, Atsushi
• Primary Examiner(s): Kinkead, Arnold
• Assistant Examiner(s): Tan, Richard
• Application Number:
  US12/358,863
• Publication Number:
  US 20090190821A1
• Time in Patent Office: 2,006 Days
• Field of Search: 382/133, 382/165, 382/172, 382/128, 706/20
• US Class Current: 382/133
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 2207/30092 Stomach; Gastric
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume
• CPC Class Codes: G06V 20/69 Microscopic objects, e.g. biological cells or cellular parts

Abstract

A pathological diagnosis support apparatus into which a digital color image showing a stained tissue sample is input, the apparatus including: a display that performs a display operation; and an image processor that when the digital color image is input, extracts cell nucleus areas, cytoplasm areas and glandular cavity areas, respectively, based on luminance values of pixels of the digital color image, measures basic feature quantities representing features of shapes of the respective cell nucleus areas, cytoplasm areas and glandular cavity areas, which have been extracted, determines whether or not a particular kind of area that appears in a limited case according to the disease state of the tissue sample exists, for each of the cell nucleus areas, the cytoplasm areas and the glandular cavity areas based on the luminance values and the basic feature quantities, and measures structure feature quantities representing ways in which the cell nucleus areas are positioned, based on the basic feature quantities of the cell nucleus areas, and if the particular kind of area exists, extracts the particular kind of area and makes the display show it together with the structure feature quantities and the cell nucleus areas.

Claims

1. A pathological diagnosis support apparatus to which a digital color image showing a stained tissue sample is input, the apparatus comprising:

a display that performs a display operation; and

an image processor that when the digital color image is input, extracts cell nucleus areas, cytoplasm areas and glandular cavity areas, respectively, based on luminance values of pixels of the digital color image, measures basic feature quantities representing features of shapes of the respective cell nucleus areas, cytoplasm areas and glandular cavity areas, which have been extracted, determines whether or not a particular kind of area that appears in a limited case according to the disease state of the tissue sample exists, for each of the cell nucleus areas, the cytoplasm areas and the glandular cavity areas based on the luminance values and the basic feature quantities, and measures structure feature quantities representing the ways in which the cell nucleus areas are positioned, based on the basic feature quantities of the cell nucleus areas, and if the particular kind of area exists, extracts the particular kind of area and makes the display show it together with the structure feature quantities and the cell nucleus areas,

wherein the image processor includes;

a memory that stores the digital color image, the basic feature quantities, and the structure feature quantities;

a feature measuring unit that measures for each of the cell nucleus areas a vector length, which is a value for characterizing each of the cell nucleus areas as having a round shape or an oval shape, an angle value representing a direction of a major axis of each of the cell nucleus areas in relation to an angle, and alignment, of cell nuclei, including thickness of in gland ducts, as the basic feature quantities; and

a display controller that makes the display show the cell nucleus areas in colors predetermined in relation to the vector lengths and the angle values,

wherein cancer cells are identified based upon a manner in which nuclei of the cancer cells are positioned, polarities of the nuclei corresponding to the manner, by measuring the polarities, the polarities representing variations in area and/or shape of the nuclei from tissue inner sides to tissue outer sides,

wherein the cancer cells are detected as cancerous due to the polarities not being in a fixed direction and being disordered, and wherein other cells are not detected as cancerous where polarities of nuclei of the other cells are in the fixed direction and are ordered.

2. A method for supporting a pathological diagnosis using a digital color image showing a stained tissue sample, the method comprising:

extracting cell nucleus areas, cytoplasm areas and glandular cavity areas, respectively, based on luminance values of pixels of the digital color image;

measuring basic feature quantities representing features of shapes of the respective cell nucleus areas, cytoplasm areas and glandular cavity areas;

determining whether or not a particular kind of area that appears in a limited case according to the disease state of the tissue sample exists, for each of the cell nucleus areas, the cytoplasm areas and the glandular cavity areas based on the luminance values and the basic feature quantities;

measuring structure feature quantities representing ways in which the cell nucleus areas are positioned, based on the basic feature quantities of the cell nucleus areas;

if the particular kind of area exists, extracting the particular kind of area and displaying it together with the structure feature quantities and the cell nucleus areas;

measuring for each of the cell nucleus areas a vector length, which is a value for characterizing each of the cell nucleus areas as having a round shape or an oval shape, an angle value representing a direction of a major axis of each of the cell nucleus areas in relation to an angle, and alignment, of cell nuclei, including thickness of in gland ducts, as the basic feature quantities; and

displaying the cell nucleus areas in colors predetermined in relation to the vector lengths and the angle values,

wherein cancer cells are identified based upon a manner in which nuclei of the cancer cells are positioned, polarities of the nuclei corresponding to the manner, by measuring the polarities, the polarities representing variations in area and/or shape of the nuclei from tissue inner sides to tissue outer sides,

wherein the cancer cells are detected as cancerous due to the polarities not being in a fixed direction and being disordered, and wherein other cells are not detected as cancerous where polarities of nuclei of the other cells are in the fixed direction and are ordered.

3. A non-transitory computer readable medium storing a program for causing a computer to execute processing for a digital color image showing a stained tissue sample, the processing comprising:

extracting cell nucleus areas, cytoplasm areas and glandular cavity areas, respectively, based on luminance values of pixels of the digital color image;

measuring basic feature quantities representing features of shapes of the respective cell nucleus areas, cytoplasm areas and glandular cavity areas, which have been extracted;

determining whether or not a particular kind of area that appears in a limited number of cases according to the disease state of the tissue sample exists, for each of the cell nucleus areas, the cytoplasm areas and the glandular cavity areas based on the luminance values and the basic feature quantities;

measuring structure feature quantities representing ways in which the cell nucleus areas are positioned, based on the basic feature quantities of the cell nucleus areas;

if the particular kind of area exists, extracting the particular kind of area and displaying it together with the structure feature quantities and the cell nucleus areas;

measuring for each of the cell nucleus areas a vector length, which is a value for characterizing each of the cell nucleus areas as having a round shape or an oval shape, an angle value representing a direction of a major axis of each of the cell nucleus areas in relation to an angle, and alignment, of cell nuclei, including thickness of in gland ducts, as the basic feature quantities; and

displaying the cell nucleus areas in colors predetermined in relation to the vector lengths and the angle values,

wherein cancer cells are identified based upon a manner in which nuclei of the cancer cells are positioned, polarities of the nuclei corresponding to the manner, by measuring the polarities, the polarities representing variations in area and/or shape of the nuclei from tissue inner sides to tissue outer sides,

wherein the cancer cells are detected as cancerous due to the polarities not being in a fixed direction and being disordered, and wherein other cells are not detected as cancerous where polarities of nuclei of the other cells are in the fixed direction and are ordered.