KR101489163B1 - Automated cervical cancer diagnosis method and system - Google Patents

Abstract

Disclosed are a method and a system for automatically diagnosing a cervical cancer. The method for automatically diagnosing a cervical cancer comprises the steps of: storing shape data of different cervix cell areas as reference data in advance for determining a cervical cancer according to the presence and absence of a cervical cancer and a degree of progression thereof; when an image of cervical cells collected from an inspected person is captured, pre-processing the captured image; extracting an image of a cervix cell area from the pre-processed image; and determining a cervical cancer with respect to the extracted image of the cervix cell area on the basis of the stored reference data, wherein the reference data includes at least one among data regarding sizes of a nucleus of cervix, a degree of irregularity of a cell nucleus, a position of a cell nucleus, and a ratio between a cell nucleus and a cytoplasm according to the presence and absence of a cervical cancer and a degree of progression thereof. According to the present invention, a cervical cancer can be automatically diagnosed promptly regardless of reading capability of an inspector.

Description

[0001] The present invention relates to a method and system for automatically diagnosing cervical cancer,

The present invention relates to a method and system for automatically diagnosing cervical cancer, and more particularly, to a method and system for automatically diagnosing cervical cancer that can be automatically diagnosed irrespective of an examiner's reading ability.

Cervical cancer is the second most common cancer in women worldwide, with about 80% of cervical cancer occurring in developing countries such as Asia, South America, and Africa. Cervical cancer is the fourth most common cancer in Korea, and according to the report of the Central Cancer Registry (1999 ~ 2002), among the 46,476 women with CAL, there were 4,394 cases of cervical cancer, accounting for 9.5% have.

Cervical cancer can be cured by early diagnosis and treatment, but when the disease progresses, the cure rate is greatly reduced according to the degree of spread. Therefore, women need regular cervical cancer screening.

Cervical cancer screening includes Pap test and liquid baseline cytology. These test methods are methods of collecting cervical cells and directly observing cancer cells by a microscope, and there is a risk that the result may vary according to the ability of the examinee to read, and it takes a lot of time.

It is an object of the present invention to provide a method and system for automatic diagnosis of cervical cancer which can automatically diagnose cervical cancer quickly regardless of the ability of a tester to read.

The technical problems of the present invention are not limited to the above-mentioned problems, and other technical problems which are not mentioned can be clearly understood by those skilled in the art from the following description.

A method for automatically diagnosing cervical cancer according to an embodiment of the present invention includes:

Storing shape data of different cervical cell regions in advance as reference data for discriminating cervical cancer according to the presence or absence and extent of cervical cancer; A step of preprocessing the captured image when the cervix cells taken from the examinee are picked up; Extracting an image of the cervix cell region from the preprocessed image; And discriminating cervical cancer for an image of the extracted cervical cell region based on the stored reference data, wherein the reference data includes at least one of a nuclear size of the different cervical cells of the different cervical cells according to the presence or absence of the cervical cancer, , Irregularity of the cell nucleus, location of the cell nucleus, and data on the ratio of the cell nucleus and the cytoplasm.

Data on the nucleus size of different cervical cells according to the presence or absence of the cervical cancer and the progression level of the cervical cancer are in the range of 15 to 25 mu m at the steady state, 25 to 30 mu m at the ASCUS (Atypical Squamous Cells of Undetermined Significance) -quadrature Squamous Intraepithelial Lesion), 38 to 45 占 퐉 for HSIL (High-grade Squamous Intraepithelial Lesion), and 15 占 퐉 or less for HPV (Human Papillomavirus).

The preprocessing may include extracting an RGB image from the captured image; Converting the RGB image into a gray image; Removing noise of the gray image; And detecting an edge of the noise-removed image.

The step of removing the noise may use an average mask.

The step of detecting the edge may be performed by any one of Sobel, Roberts, Laplacian, and Canny edge detection methods.

The step of extracting the image of the cervical cell region may be performed by any one of a Hough Transform, a region of interest (ROI), and a watershed using pixels of an image.

Further, in the automatic diagnosis system for cervical cancer according to the embodiment of the present invention,

A database for storing shape data of different cervical cell regions as reference data for discriminating cervical cancer according to the presence and progression of cervical cancer; And extracting an image of a cervix cell region from the preprocessed image, and based on the stored reference data, extracting an image of the extracted cervical cell region Wherein the reference data includes at least one of a nucleus size of the different cervical cells, an irregularity of the cell nucleus, a location of the cell nucleus, and a cell nucleus according to the presence or absence of the cervical cancer and the degree of progression of the cervical cancer. And data on the ratio of cytoplasm.

Data on the nucleus size of different cervical cells according to the presence or absence of the cervical cancer and the progression level of the cervical cancer are in the range of 15 to 25 mu m at the steady state, 25 to 30 mu m at the ASCUS (Atypical Squamous Cells of Undetermined Significance) -quadrature Squamous Intraepithelial Lesion), 38 to 45 占 퐉 for HSIL (High-grade Squamous Intraepithelial Lesion), and 15 占 퐉 or less for HPV (Human Papillomavirus).

Wherein the diagnostic apparatus extracts an RGB image from the sensed image, converts the RGB image to a gray image, removes noise of the gray image, prepares the noise- It is possible to detect an edge.

The diagnostic device may use an average mask to remove the noise.

The diagnostic apparatus can use any one of Sobel, Roberts, Laplacian, and Canny edge detection methods when detecting the edge.

The diagnostic apparatus may use any one of Hough Transform, Region of Interest (ROI), and watershed using pixels of an image when extracting the image of the cervical cell region.

The method and system for automatically diagnosing cervical cancer according to the present invention have the effect of automatically diagnosing cervical cancer promptly regardless of the ability of the examinee to read.

1 is a block diagram of a system for automatically diagnosing cervical cancer according to an embodiment of the present invention.
2 is a diagram showing an example of a database screen in which reference data is stored.
3 is a diagram showing an example of reference data.
4 is a flowchart illustrating a method for automatically diagnosing cervical cancer according to an embodiment of the present invention.

Brief Description of the Drawings The advantages and features of the present invention and methods of accomplishing the same will become apparent by reference to the embodiments described in detail below with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. To fully disclose the scope of the invention to those skilled in the art, and the present invention is only defined by the claims. Like reference numerals refer to like elements throughout the specification.

Hereinafter, a method and system for automatically diagnosing cervical cancer according to an embodiment of the present invention will be described with reference to the accompanying drawings.

FIG. 1 is a block diagram of a system for automatically diagnosing cervical cancer according to an embodiment of the present invention. FIG. 2 is a view showing an example of a database screen in which reference data is stored, and FIG. 3 is an example of reference data .

Referring to FIG. 1, a cervical cancer auto-diagnosis system 100 according to an embodiment of the present invention includes a database 110, an imaging device 130, and a diagnostic device 150.

In the database 110, shape data of different cervical cell regions are stored as reference data for discriminating cervical cancer according to the presence or absence and extent of cervical cancer. This database 110 is created and managed in software.

The region of the cervical cell may be a nucleus region, a cytoplasmic region, or a whole cell region including a nucleus region and a cytoplasmic region. However, in the present embodiment, it is assumed that the region is a nucleus region.

The reference data may include at least one of the nucleus size of the cervix cells, the degree of irregularity of the cell nucleus, the location of the cell nucleus, and the data on the ratio of the cell nucleus and the cytoplasm.

As shown in FIG. 2, the reference data are classified into a normal state, an ASCII (Atypical Squamous Cells of Undetermined Significance), a low-grade squamous intraepithelial lesion (LSIL), a high-grade squamous Intraepithelial Lesion, and HPV (Human Papillomavirus). The reference data for each state may be stored as image data as shown in FIG. 3, or may be stored as character data as shown in Table 1 below.

When the reference data includes data on the size of the cervical cell nuclei, the data on the nucleus size of the cervical cells are shown in Table 1, depending on the presence and progression of the cervical cancer, 25 to 30 占 퐉 when ASCUS (Atypical Squamous Cells of Undetermined Significance), 30 to 38 占 퐉 when LSIL (Low-grade Squamous Intraepithelial Lesion), 38 to 45 占 퐉 when HSIL (High-grade Squamous Intraepithelial Lesion) (Human Papillomavirus).

Condition Cell nucleus size Diagnosis One 15 to 25 μm Normal 2 25 to 30 μm ASCUS 3 30 ~ 38㎛ LSIL 4 38 to 45 μm HSIL 5 15 μm or less HPV

The imaging device 130 images the cervical cells collected from the examinee. Such an imaging device 130 may include a microscope.

The diagnostic apparatus 150 can receive an image captured by the image capturing apparatus 130 by being electrically connected to the image capturing apparatus 130. The diagnostic device 150 receives and preprocesses an image captured by the imaging device 130, and extracts an image of a cervix cell region (for example, a nucleus region) from the preprocessed image. Based on the reference data stored in the database 110, the diagnostic device 150 determines cervical cancer for the image of the extracted cervical cell area. The operation of the diagnostic device 150 is stored in a programmed state in the internal memory and is executed at the time of diagnosis.

The diagnostic apparatus 150 extracts an RGB image from the captured image when preprocessing the image captured by the image capturing apparatus 130. The diagnostic apparatus 150 converts the RGB image into a gray image, removes the noise of the gray image, and detects an edge of the noise-removed image. This image preprocessing process can be performed using Labview's Vision Development, Matlab, or a combination of LabVIEW and MATLAB.

Functions such as an RGB threshold, a gray threshold, a morphology, a Gaussian filter, and a convolution filter can be used in image preprocessing using LabVIEW.

During image preprocessing using MATLAB, the extracted RGB image is applied to Equation (1).

[Equation 1]

The RGB image obtained by Equation (1) is converted into a gray image, and the converted gray image can be removed by an average mask (for example, a 3 * 3 mask). The noise canceled image can be detected by any of the Sobel, Roberts, Laplacian, and Canny edge detection methods.

The diagnostic device 150 may use Hough Transform, Region of Interest (ROI), and pixels of the image when extracting an image of a cervical cell region (e.g., a nucleus region) from a preprocessed image Any one of water sheds may be used. The extracted cervical cell area may be marked with a circular line so that it is distinct from other areas.

When the reference data is as shown in Table 1, the diagnosis apparatus 150 determines whether the size of the nucleus region as the extracted cervical cell region is 15 to 25 탆, the normal state, 25 to 30 탆 ASCUS, 30 to 38 탆 LSIL, 38 To 45 탆, and HPV if it is 15 탆 or less. If the reference data is stored as image data, the diagnostic apparatus 150 can discriminate the cervical cancer according to whether the size of the nucleus region of the extracted cervical cells matches the size of the nucleus region of the image data. Then, the diagnostic apparatus 150 can display the result of the discrimination of cervical cancer on the screen.

A method for automatically diagnosing cervical cancer using the system 100 will be described with reference to FIG.

Referring to FIG. 4, first, shape data of different cervical cells are stored in advance in the database 110 as reference data for discriminating cervical cancer according to the presence or absence and extent of cervical cancer (S110).

When the cervical cells collected from the examinee are picked up by the image pickup device 130, the diagnostic device 150 receives the picked-up image (S120). The diagnostic apparatus 150 preprocesses the input image. In the preprocessing step, the RGB image is extracted from the captured image (S130), the extracted RGB image is converted into a gray image (S140), the noise of the gray image is removed using an average mask (S150) The edge of the image is detected by any one of Sobel, Roberts, Laplacian, and Canny methods (S160).

The diagnostic apparatus 150 extracts an image of the cervical cell region from the preprocessed image (S170). Image extraction of the cervical cell region can be performed by either Hough Transform, Region of Interest (ROI), or watershed using pixels of an image.

Finally, based on the reference data stored in the database 110, the diagnostic apparatus 150 identifies the cervical cancer for the image of the extracted cervical cell region and displays the result on the screen (S180).

Meanwhile, the embodiments of the present invention described above can be implemented in a general-purpose digital computer that can be created as a program that can be executed by a computer and operates the program using a computer-readable recording medium.

The computer readable recording medium may be a magnetic storage medium such as a ROM, a floppy disk, a hard disk, etc., an optical reading medium such as a CD-ROM or a DVD and a carrier wave such as the Internet Lt; / RTI > transmission).

The present invention has been described with reference to preferred embodiments thereof. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

As described above, according to the method and system for automatically diagnosing cervical cancer according to the present invention, it is possible to automatically diagnose cervical cancer promptly regardless of the ability of the examinee to read.

100: Auto diagnosis system of cervical cancer 110: Database
130: Image pickup device 150: Diagnostic device

Claims (14)

Storing shape data of different cervical cell regions in advance as reference data for discriminating cervical cancer according to the presence or absence and extent of cervical cancer;
Pre-processing the captured image when the cervix cells taken from the examinee are picked up;
Extracting an image of the cervix cell region from the preprocessed image; And
And discriminating cervical cancer for an image of the extracted cervical cell region based on the stored reference data,
Wherein the reference data includes at least one of data on the nucleus size of the different cervical cells, the degree of irregularity of the cell nucleus, the location of the cell nucleus, and the ratio of the cell nucleus and the cytoplasm in accordance with the presence or absence of the cervical cancer, Method for automatic diagnosis of cervical cancer.
delete The method according to claim 1,
Data on the nucleus size of different cervical cells according to the presence or absence of the cervical cancer and the progression level of the cervical cancer are in the range of 15 to 25 mu m at the steady state, 25 to 30 mu m at the ASCUS (Atypical Squamous Cells of Undetermined Significance) -aggress Squamous Intraepithelial Lesion), 38-45 [mu] m at HSIL (High-grade Squamous Intraepithelial Lesion), and 15 [mu] m at HPV (Human Papillomavirus).
The method according to claim 1,
The pre-
Extracting an RGB image from the captured image;
Converting the RGB image into a gray image;
Removing noise of the gray image; And
And detecting an edge of the noise-removed image.
5. The method of claim 4,
The method of claim 1, wherein the noise is removed using an average mask.
5. The method of claim 4,
Wherein the step of detecting the edge is performed by any one of Sobel, Roberts, Laplacian, and Canny edge detection methods.
The method according to claim 1,
Wherein the step of extracting the image of the cervical cell region is performed by any one of a Hough Transform, a Region of Interest (ROI), and a watershed using pixels of an image.
A database for storing shape data of different cervical cell regions as reference data for discriminating cervical cancer according to the presence and progression of cervical cancer; And
The image processing method according to claim 1, wherein, when the cervical cells taken from the subject are picked up, the captured image is preprocessed, an image of the cervix cell region is extracted from the preprocessed image, and based on the stored reference data, And a diagnostic device for identifying the cervical cancer of the cervix,
Wherein the reference data includes at least one of data on the nucleus size of the different cervical cells, the degree of irregularity of the cell nucleus, the location of the cell nucleus, and the ratio of the cell nucleus and the cytoplasm in accordance with the presence or absence of the cervical cancer, Automated Diagnosis System for Cervical Cancer.
delete 9. The method of claim 8,
Data on the nucleus size of different cervical cells according to the presence or absence of the cervical cancer and the progression level of the cervical cancer are in the range of 15 to 25 mu m at the steady state, 25 to 30 mu m at the ASCUS (Atypical Squamous Cells of Undetermined Significance) -grade SQUAMOUS Intraepithelial Lesion), 38-45 μm for HSIL (High-grade Squamous Intraepithelial Lesion), and 15 μm for HPV (Human Papillomavirus).
9. The method of claim 8,
Wherein the diagnostic apparatus extracts an RGB image from the sensed image, converts the RGB image to a gray image, removes noise of the gray image, prepares the noise- Automatic diagnosis system of cervical cancer detecting edge.
12. The method of claim 11,
Wherein the diagnostic apparatus uses an average mask to remove the noise.
12. The method of claim 11,
Wherein the diagnostic apparatus uses any one of Sobel, Roberts, Laplacian, and Canny edge detection methods when detecting the edge.
9. The method of claim 8,
The diagnostic apparatus includes a cervical cancer auto diagnosis system using any one of a Hough Transform, a Region of Interest (ROI), and a watershed using pixels of an image in extracting an image of the cervical cell region .

Images