CN110992303B - Abnormal cell screening method and device, electronic equipment and storage medium - Google Patents
Abnormal cell screening method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN110992303B CN110992303B CN201911040955.9A CN201911040955A CN110992303B CN 110992303 B CN110992303 B CN 110992303B CN 201911040955 A CN201911040955 A CN 201911040955A CN 110992303 B CN110992303 B CN 110992303B
- Authority
- CN
- China
- Prior art keywords
- cervical cell
- image
- images
- cell sub
- cervical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 147
- 238000012216 screening Methods 0.000 title claims abstract description 93
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000003860 storage Methods 0.000 title claims abstract description 9
- 238000013528 artificial neural network Methods 0.000 claims abstract description 77
- 238000012549 training Methods 0.000 claims description 49
- 230000011218 segmentation Effects 0.000 claims description 46
- 238000009792 diffusion process Methods 0.000 claims description 14
- 230000001788 irregular Effects 0.000 claims description 14
- 230000003902 lesion Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000004891 communication Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 4
- 231100000915 pathological change Toxicity 0.000 claims description 2
- 230000036285 pathological change Effects 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 abstract 1
- 239000003550 marker Substances 0.000 description 14
- 208000007879 Atypical Squamous Cells of the Cervix Diseases 0.000 description 10
- 206010008263 Cervical dysplasia Diseases 0.000 description 10
- 208000009956 adenocarcinoma Diseases 0.000 description 10
- 238000010586 diagram Methods 0.000 description 10
- 230000009841 epithelial lesion Effects 0.000 description 8
- 206010041823 squamous cell carcinoma Diseases 0.000 description 6
- 206010008342 Cervix carcinoma Diseases 0.000 description 4
- 208000032124 Squamous Intraepithelial Lesions Diseases 0.000 description 4
- 208000006105 Uterine Cervical Neoplasms Diseases 0.000 description 4
- 201000010881 cervical cancer Diseases 0.000 description 4
- 206010028980 Neoplasm Diseases 0.000 description 3
- 201000011510 cancer Diseases 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 238000005520 cutting process Methods 0.000 description 2
- 210000000981 epithelium Anatomy 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 230000034994 death Effects 0.000 description 1
- 231100000517 death Toxicity 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000036210 malignancy Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Radiology & Medical Imaging (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Pathology (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The application relates to a neural network and discloses an abnormal cell screening method, an abnormal cell screening device, electronic equipment and a storage medium, wherein the abnormal cell screening method comprises the following steps: obtaining a cervical cell image; segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images; respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images; selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results; acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results; displaying the plurality of first cervical cell sub-images on a display interface. The embodiment of the invention is beneficial to improving the screening efficiency of abnormal cells.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method and apparatus for screening abnormal cells, an electronic device, and a storage medium.
Background
Cervical cancer is the most common gynaecological malignancy, and in recent years the incidence has a trend to lower the age, with 50 tens of thousands of new cases and 27.4 dead cases worldwide each year, with 85% of cervical cancer deaths occurring in low-medium income areas with low census rates. Cervical cancer is the only cancer which can be found and cured at present, so that early screening and diagnosis are key links for preventing and treating cervical cancer.
Currently, in general physical examination centers and hospitals, doctors are generally required to find abnormal cells from thousands of cells under a microscope and to diagnose according to the abnormal cells. This mode of abnormal cell screening is inefficient.
Disclosure of Invention
The embodiment of the invention provides an abnormal cell screening method, an abnormal cell screening device, electronic equipment and a storage medium.
The first aspect of the present invention provides a method for screening abnormal cells, comprising:
obtaining a cervical cell image;
segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells;
Selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results;
acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
displaying the plurality of first cervical cell sub-images on a display interface;
the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells.
In a second aspect, the present invention provides an abnormal cell screening apparatus comprising:
the first acquisition module is used for acquiring cervical cell images;
the segmentation module is used for segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
the input module is used for respectively inputting the cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with pathological changes or cancerations on the basis of the cervical cells;
The selecting module is used for selecting a plurality of predicted results from the plurality of predicted results according to a preset selecting strategy to serve as a plurality of first predicted results;
the second acquisition module is used for acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
a display module for displaying the plurality of first cervical cell sub-images on a display interface;
the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells.
A third aspect of the present invention provides an electronic device for abnormal cell screening comprising a processor, a memory, a communication interface and one or more programs, wherein the one or more programs are stored in the memory and are generated for execution by the processor to perform instructions for steps in any one of the methods of an abnormal cell screening method.
A fourth aspect of the present invention provides a computer readable storage medium for storing a computer program for execution by the processor to implement the method of any one of the abnormal cell screening methods.
It can be seen that in the above technical solution, cervical cell images are acquired; segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images; respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells; selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results; acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results; displaying the plurality of first cervical cell sub-images on a display interface. By segmenting the cervical cell image into a plurality of cervical cell sub-images and respectively inputting the plurality of cervical cell sub-images into the abnormal cell screening model, the condition that the screening process consumes much time due to the fact that one cervical cell image is too large is avoided, and the abnormal cell screening model screening efficiency is accelerated. Meanwhile, the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, the plurality of neural networks are used for screening different abnormal cells, and therefore a plurality of abnormal cells can be screened out, and the practicability is higher. Furthermore, the plurality of first prediction results with the highest prediction probability are screened out, and the plurality of first cervical cell sub-images corresponding to the plurality of first prediction results are displayed on the interface, so that the doctor can more conveniently check the images, and the workload of the doctor is reduced by displaying the cervical cell sub-image with the highest abnormal cell probability to the doctor.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Wherein:
FIG. 1A is a flow chart of an abnormal cell screening method according to an embodiment of the present invention;
FIG. 1B is a schematic diagram of a display interface according to an embodiment of the present invention;
FIG. 2A is a flow chart of another method for screening abnormal cells according to an embodiment of the present invention;
FIG. 2B is a schematic diagram of a coordinate system according to an embodiment of the present invention;
FIG. 2C is a schematic illustration of a cut shape provided by an embodiment of the present invention;
FIG. 2D is a schematic illustration of pixel diffusion according to an embodiment of the present invention;
FIG. 3 is a flow chart of another method for screening abnormal cells according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an abnormal cell screening apparatus according to an embodiment of the present invention;
Fig. 5 is a schematic diagram of a server structure of a hardware running environment according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The following will describe in detail.
The terms first and second in the description and claims of the invention and in the above-mentioned figures are used for distinguishing between different objects and not for describing a particular sequential order. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.
First, the execution body in the embodiment of the present application may be, for example, a server, or may be a local data processing device. The server may be, for example, a tablet computer, a notebook computer, a palm computer, an MID, a desktop computer, or other server devices. No limitation is made in the present application.
101. Obtaining a cervical cell image;
optionally, the acquiring cervical cell image includes: and acquiring the cervical cell image through a scanning device.
The scanning device may be, for example, a scanner. It will be appreciated that a scanner is used to scan the cervical cell layer detected by the liquid-based lamellar cells to obtain a cervical cell image.
Wherein the cervical cell image is an image of a cervical cell layer.
102. Segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
103. respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells;
Wherein, the cervical cells are cells in a normal growth state.
The abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells.
Further, a first neural network is one of the plurality of neural networks, the first neural network being for screening for first abnormal cells included in each of the plurality of cervical cell sub-images, and a second neural network is another of the plurality of neural networks different from the first neural network, the second neural network being for screening for second abnormal cells included in each of the plurality of cervical cell sub-images.
Wherein the first abnormal cell is different from the second abnormal cell, the first abnormal cell is one of the abnormal cells, and the second abnormal cell is another of the abnormal cells.
Alternatively, the abnormal cells include at least one of: squamous carcinoma (squamous cell carcinom, SCC), high-grade squamous epithelial lesions (high-grade squamous epithelial lesion, LSIL), atypical squamous cells that cannot exclude high-grade squamous intraepithelial lesions (atypical squamous cells, cannot exclude high-gradesquamous intraepithelial lesion, ASC-H))), low-grade squamous epithelial lesions (low-grade squamous epithelial lesion, LSIL), atypical squamous cells of undefined significance (atypical squamouscells of unde. Fixed signature, ASC-US), adenocarcinoma (AC), atypical adenocarcinoma (Atypical adenocarcinoma, AGC), and the like.
It will be appreciated that the plurality of neural networks are used to screen each of the plurality of cervical cell sub-images for an abnormal cell. Specifically, a certain neural network of the plurality of neural networks is used to screen squamous cell carcinoma included in each of the plurality of cervical cell sub-images, and another neural network of the plurality of neural networks is used to screen adenocarcinoma included in each of the plurality of cervical cell sub-images.
104. Selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results;
the preset selection strategy is determined according to a prediction selection operation, and the prediction selection operation comprises the following steps: obtaining a prediction probability corresponding to each prediction result in the plurality of prediction results; numbering the prediction probability corresponding to each of the plurality of prediction results according to the order of the prediction probabilities from large to small so as to obtain a plurality of numbers; selecting a part number from the plurality of numbers, wherein the part number is at least one number which is larger than a preset number in the plurality of numbers; and taking the part prediction probability corresponding to the part numbers one by one as the preset selection strategy.
Wherein the preset number is set by an administrator. For example, the number is 10 numbers from 10-1, the preset number is 6, then the partial number is 3 numbers of 7-10.
Further, the selecting a plurality of predicted results from the plurality of predicted results according to a preset selection policy as a plurality of first predicted results includes: and selecting a plurality of prediction results from the plurality of prediction results according to the partial prediction probability as a plurality of first prediction results.
105. Acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
106. displaying the plurality of first cervical cell sub-images on a display interface.
Referring to fig. 1B, fig. 1B is a schematic diagram of a display interface provided in an embodiment of the present invention, where each of the rectangles with diagonal lines represents a first cervical cell sub-image. It can be seen that a plurality of first cervical cell sub-images are displayed on the display interface.
It can be seen that in the above technical solution, cervical cell images are acquired; segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images; respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells; selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results; acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results; displaying the plurality of first cervical cell sub-images on a display interface. By segmenting the cervical cell image into a plurality of cervical cell sub-images and respectively inputting the plurality of cervical cell sub-images into the abnormal cell screening model, the condition that the screening process consumes much time due to the fact that one cervical cell image is too large is avoided, and the abnormal cell screening model screening efficiency is accelerated. Meanwhile, the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, the plurality of neural networks are used for screening different abnormal cells, and therefore a plurality of abnormal cells can be screened out, and the practicability is higher. Furthermore, the plurality of first prediction results with the highest prediction probability are screened out, and the plurality of first cervical cell sub-images corresponding to the plurality of first prediction results are displayed on the interface, so that the doctor can more conveniently check the images, and the workload of the doctor is reduced by displaying the cervical cell sub-image with the highest abnormal cell probability to the doctor.
This process of segmenting the cervical cell image from the gray values based on pixels in the cervical cell image to obtain a plurality of cervical cell sub-images is specifically illustrated below.
Referring to fig. 2A, fig. 2A is a schematic flow chart of an abnormal cell screening method according to an embodiment of the invention. As shown in fig. 2A, an abnormal cell screening method according to an embodiment of the present invention may include:
201. acquiring a gray value corresponding to each pixel point in the cervical cell image;
wherein the gray value ranges from 0 to 255.
When the cervical cell image includes abnormal cells, the abnormal cells are different, and the corresponding gray values are also different. For example, when the cervical cell image includes squamous carcinoma and highly squamous epithelium lesions, the gradation value corresponding to the squamous carcinoma is different from the gradation value corresponding to the highly squamous epithelium lesions.
202. Determining a pixel point with the minimum gray value according to the gray value corresponding to each pixel point in the cervical cell image;
203. selecting any one pixel point from the pixel points as a coordinate origin for dividing the cervical cell image;
204. establishing a coordinate system on the cervical cell image based on the origin of coordinates, wherein the coordinate system takes the transverse positive direction of the cervical cell image as an x-axis and the longitudinal positive direction of the cervical cell image as a y-axis;
Referring to fig. 2B, fig. 2B is a schematic diagram of a coordinate system according to an embodiment of the present invention, it can be seen that the positive direction of the x-axis of the coordinate system is the right direction, that is, the transverse positive direction of the cervical cell image; the positive direction of the y-axis of the coordinate system is the upward direction, namely the longitudinal positive direction of the cervical cell image; the coordinate system is established on the cervical cell image based on the origin of coordinates.
205. And cutting the cervical cell image from the origin of coordinates to obtain the plurality of cervical cell sub-images.
Optionally, in one possible embodiment, the segmenting the cervical cell image from the origin of coordinates to obtain the plurality of cervical cell sub-images includes: determining a plurality of gray difference values according to gray values corresponding to each pixel point in the cervical cell image, wherein each gray difference value is a difference value between each pixel point and a corresponding adjacent pixel point on the gray value; dividing the gray difference values into a group which falls into the same gray value interval to obtain a plurality of gray groups; determining a plurality of segmentation shapes corresponding to the cervical cell image from the origin of coordinates according to the plurality of gray groups; normalizing the irregular shape in the plurality of segmentation shapes to obtain a plurality of first segmentation shapes with rules; the plurality of first cut-out shapes are set as the plurality of cervical cell sub-images.
Optionally, in one possible implementation manner, the first pixel is any one pixel in the cervical cell image, and the determining a plurality of gray scale differences according to gray scale values corresponding to each pixel in the cervical cell image includes: acquiring at least one gray value corresponding to at least one pixel point adjacent to the first pixel point; and respectively determining the gray level difference value between the gray level value corresponding to the first pixel point and each gray level value in the at least one gray level value to obtain at least one gray level difference value.
For example, the cervical cell image includes 9 pixels, the 9 pixels are arranged in a square, and when a pixel at the center of the square is a first pixel, at least one pixel adjacent to the first pixel includes: adjacent pixel points above the first pixel point, adjacent pixel points below the first pixel point, adjacent pixel points to the left of the first pixel point and adjacent pixel points to the right of the first pixel point; when the pixel point at the top right corner of the square is the first pixel point, at least one pixel point adjacent to the first pixel point comprises: adjacent pixel points below the first pixel point, adjacent pixel points to the left of the first pixel point and adjacent pixel points to the right of the first pixel point.
Further, at least one pixel point adjacent to the first pixel point at least includes one of the following: adjacent pixel points above the first pixel point, adjacent pixel points below the first pixel point, adjacent pixel points to the left of the first pixel point and adjacent pixel points to the right of the first pixel point.
It can be seen that, in the above technical solution, at least one gray value corresponding to at least one pixel point adjacent to the first pixel point is obtained; and respectively determining the gray level difference value between the gray level value corresponding to the first pixel point and each gray level value in the at least one gray level value to obtain at least one gray level difference value, so as to determine the difference value of the adjacent pixel points corresponding to the first pixel point on the gray level value, and prepare for grouping according to the same gray level value interval in which the gray level difference value falls.
It will be appreciated that when determining a plurality of segmentation shapes corresponding to the cervical cell image from the origin of coordinates according to the plurality of gray groups, it is highly probable that the segmentation shapes are not regular shapes. Among them, the regular shape includes, for example: rectangular, square, etc. Therefore, the irregular shapes in the plurality of segmentation shapes need to be normalized to obtain the regular first segmentation shapes, so that the abnormal cells can be screened out by the abnormal cell screening model more quickly.
Wherein the irregular shape is a shape excluding a rectangle and a square, and the normalization process includes the operations of: determining a plurality of second pixel points corresponding to each segmentation shape in the plurality of segmentation shapes, wherein the gray value corresponding to each second pixel point is a first gray value; and carrying out preset processing on a plurality of second pixel points corresponding to each segmentation shape in the plurality of segmentation shapes according to a first sequence.
Wherein the first order comprises one of: the center point of each of the plurality of segmented shapes is in a near-to-far order from the origin of coordinates and the center point of each of the plurality of segmented shapes is in a far-to-near order from the origin of coordinates.
Further, the preset process includes at least one of the following processes: removing redundant second pixel points in each segmentation shape; collecting redundant second pixel points in the plurality of segmentation shapes to obtain a plurality of third pixel points, and filling the segmentation shapes needing filling in the plurality of segmentation shapes by adopting the plurality of third pixel points; and rearranging the plurality of second pixel points corresponding to each segmentation shape. The redundant second pixel points are determined according to the shape attribute corresponding to each segmentation shape, and the shape attribute corresponding to each segmentation shape is determined according to the process that each segmentation shape tends to be a regular shape.
For example, a certain segmentation shape of the plurality of segmentation shapes is a triangle, referring to fig. 2C, fig. 2C is a schematic diagram of a segmentation shape provided by an embodiment of the present invention, wherein an uppermost legend is to directly remove redundant second pixel points in the triangle, so as to obtain a rectangle, that is, a regular first segmentation shape; the middle illustration is that firstly redundant second pixel points in the triangle are rearranged, then redundant second pixel points in other segmentation shapes are adopted to fill the arranged triangle, and finally a rectangle is formed, namely a regular first segmentation shape; the last legend is a rectangle obtained by rearranging redundant second pixel points in the triangle, namely a first segmentation shape.
Wherein the first gray value is 0 or 255.
Optionally, in the above technical solution, a plurality of gray level differences are determined according to gray level values corresponding to each pixel point in the cervical cell image, where each gray level difference is a difference between each pixel point and a gray level value of an adjacent pixel point corresponding to each pixel point; dividing the gray difference values into a group which falls into the same gray value interval to obtain a plurality of gray groups; determining a plurality of segmentation shapes corresponding to the cervical cell image from the origin of coordinates according to the plurality of gray groups; normalizing the irregular shape in the plurality of segmentation shapes to obtain a plurality of first segmentation shapes with rules; the plurality of first segmentation shapes are set as the plurality of cervical cell sub-images, so that a plurality of segmentation shapes corresponding to the cervical cell images are determined from the origin of coordinates according to the gray scale group, each segmentation shape is composed of pixels falling into the same gray scale value interval in the gray scale difference value, the abnormal cell screening model screens abnormal cells more quickly, and the image segmentation efficiency is also accelerated. Meanwhile, irregular shapes in the plurality of segmentation shapes are normalized, so that input data more suitable for an abnormal cell screening model is constructed, and screening efficiency is quickened.
It can be seen that in one possible embodiment, the segmenting the cervical cell image based on the gray values of the pixels in the cervical cell image to obtain the plurality of cervical cell sub-images includes: acquiring a gray value corresponding to each pixel point in the cervical cell image; determining a plurality of groups of pixel points with the same gray value in the cervical cell image according to the gray value corresponding to each pixel point in the cervical cell image, wherein each group of pixel points comprises at least one pixel point, and the gray values corresponding to each pixel point in the at least one pixel point are the same; dividing the cervical cell image according to the plurality of groups of pixel points to obtain a plurality of cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of cervical cell sub-images.
According to the technical scheme, the gray value corresponding to each pixel point in the cervical cell image is obtained; determining a plurality of groups of pixel points with the same gray value in the cervical cell image according to the gray value corresponding to each pixel point in the cervical cell image, wherein each group of pixel points comprises at least one pixel point, and the gray values corresponding to each pixel point in the at least one pixel point are the same; dividing the cervical cell image according to the plurality of groups of pixel points to obtain the plurality of cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of cervical cell sub-images, so that the cervical cell image is divided according to the same gray value, and preparation is made for screening abnormal cells faster by a follow-up abnormal cell screening model.
Optionally, in one possible implementation manner, the dividing the cervical cell image according to the multiple sets of pixel points to obtain the multiple cervical cell sub-images includes: dividing the cervical cell image according to the plurality of groups of pixel points to obtain a plurality of second cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of second cervical cell sub-images; determining at least one third cervical cell sub-image having an irregular outline in the plurality of second cervical cell sub-images; performing the following operations for each of the at least one third cervical cell sub-image to obtain the plurality of cervical cell sub-images, including: determining the outer contour size corresponding to the currently processed third cervical cell sub-image; obtaining a template image according to the outer contour size corresponding to the currently processed third cervical cell sub-image so as to obtain the outer contour size corresponding to the template image; and carrying out pixel diffusion on the third cervical cell sub-image which is processed currently according to the outer contour size corresponding to the template image until the outer contour size corresponding to the third cervical cell sub-image which is processed currently is the same as the outer contour size corresponding to the template image, wherein the pixel diffusion is carried out by adopting a first gray value.
Wherein the outer contour is irregular to exclude contours with rectangular and square outer contours. Wherein the template image comprises one of: rectangular and square.
Wherein the first gray value is 0 or 255.
For example, if one of the plurality of second cervical cell sub-images is triangular, the outline corresponding to the second cervical cell sub-image is triangular, and further, the outline corresponding to the second cervical cell sub-image is irregular. It can be understood that referring to fig. 2D, fig. 2D is a schematic diagram of pixel diffusion, and when the template image corresponding to the second cervical cell sub-image is rectangular, pixel diffusion is performed on the triangle according to the outer contour dimension corresponding to the rectangle, and finally, the cervical cell sub-image corresponding to the second cervical cell sub-image is rectangular.
In the above technical solution, the cervical cell images are segmented according to the plurality of groups of pixel points to obtain a plurality of second cervical cell sub-images, where the plurality of groups of pixel points correspond to the plurality of second cervical cell sub-images; determining at least one third cervical cell sub-image having an irregular outline in the plurality of second cervical cell sub-images; performing the following operations for each of the at least one third cervical cell sub-image to obtain the plurality of cervical cell sub-images, including: determining the outer contour size corresponding to the currently processed third cervical cell sub-image; obtaining a template image according to the outer contour size corresponding to the currently processed third cervical cell sub-image so as to obtain the outer contour size corresponding to the template image; and carrying out pixel diffusion on the third cervical cell sub-image which is currently processed according to the outer contour dimension corresponding to the template image until the outer contour dimension corresponding to the third cervical cell sub-image which is currently processed is the same as the outer contour dimension corresponding to the template image, stopping the pixel diffusion, wherein the pixel diffusion adopts a first gray value for diffusion, and all the cervical cell sub-images are changed into rectangular or square images by carrying out pixel diffusion on at least one third cervical cell sub-image with irregular outer contour, so that preparation is provided for more rapid screening of abnormal cells of a follow-up abnormal cell screening model.
Referring to fig. 3, fig. 3 is a flow chart of an abnormal cell screening method according to still another embodiment of the present invention. Wherein, as shown in fig. 3, before the plurality of cervical cell images are respectively input into the abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell images, the method further includes:
301. acquiring a training set, wherein the training set comprises a plurality of training subsets, the plurality of training subsets correspond to the plurality of neural networks, each training subset comprises a plurality of cervical cell sub-image sets with different brightness degrees, each cervical cell sub-image set comprises a plurality of second cervical cell sub-images with different brightness degrees, and each second cervical cell sub-image in the plurality of second cervical cell sub-images comprises abnormal cells different;
optionally, in one possible implementation manner, the acquiring a training set includes: displaying the plurality of third cervical cell sub-images on a labeling interface; when detecting marking operations for a plurality of positions on the marking interface, marking the plurality of third cervical cell sub-images corresponding to the plurality of positions to obtain a plurality of fourth cervical cell sub-images corresponding to the plurality of third cervical cell sub-images, wherein each fourth cervical cell sub-image is an image obtained by marking each third cervical cell sub-image; processing the fourth cervical cell sub-images with a plurality of preset brightnesses respectively to obtain the cervical cell sub-image sets included in each training subset of the training subsets, wherein the preset brightnesses correspond to the cervical cell sub-image sets; setting the plurality of cervical cell sub-image sets included in each of the plurality of training subsets as the training set.
The marking interface comprises a plurality of marking display areas, and the marking display areas correspond to the third cervical cell sub-images. The displaying a plurality of third cervical cell sub-images on the label interface, comprising: the plurality of third cervical cell sub-images are displayed at the plurality of marker display areas on a marker interface. Further, a person having medical knowledge may view the plurality of third cervical cell sub-images in the plurality of marker display areas and mark the plurality of third cervical cell sub-images.
Optionally, in one possible implementation manner, when a marking operation for a plurality of positions on the marking interface is detected, marking the plurality of third cervical cell sub-images corresponding to the plurality of positions to obtain a plurality of fourth cervical cell sub-images corresponding to the plurality of third cervical cell sub-images, including: tracking a plurality of marker tracks on the plurality of third cervical cell sub-images corresponding to a plurality of locations on the marker interface when a marker operation is detected for the plurality of locations; acquiring a plurality of mark areas formed by the plurality of mark tracks; generating a plurality of marking labels according to the plurality of marking areas; generating the fourth plurality of cervical cell sub-images including the plurality of marker tags from the third plurality of cervical cell sub-images corresponding to the plurality of locations.
Wherein, the first position is any one position of the plurality of positions, and when the marking operation of the first position on the marking interface is detected, the marking track on the third cervical cell sub-image corresponding to the first position is tracked; acquiring a mark region formed by the mark track; generating a marking label according to the marking area; and generating a fourth cervical cell sub-image comprising the marking tag according to the third cervical cell sub-image corresponding to the first position.
Further, the plurality of third cervical cell sub-images correspond to the plurality of marker loci, the plurality of marker loci correspond to the plurality of marker areas, and the plurality of marker areas correspond to the plurality of marker tags.
It can be seen that in the above technical solution, when a marking operation for a plurality of positions on the marking interface is detected, a plurality of marking tracks on the plurality of third cervical cell sub-images corresponding to the plurality of positions are tracked; acquiring a plurality of mark areas formed by the plurality of mark tracks; generating a plurality of marking labels according to the plurality of marking areas; generating the fourth cervical cell sub-images including the marking labels according to the third cervical cell sub-images corresponding to the positions, wherein the marking areas are different when marking tracks are different, so that the marking labels are different, the uniqueness of the fourth cervical cell sub-image marking labels is improved, and the model training efficiency is improved.
It can be seen that in the above technical solution, a plurality of third cervical cell sub-images are displayed on the marker interface; when detecting marking operations for a plurality of positions on the marking interface, marking the plurality of third cervical cell sub-images corresponding to the plurality of positions to obtain a plurality of fourth cervical cell sub-images corresponding to the plurality of third cervical cell sub-images, wherein each fourth cervical cell sub-image is an image obtained by marking each third cervical cell sub-image; processing the fourth cervical cell sub-images with a plurality of preset brightnesses respectively to obtain the cervical cell sub-image sets included in each training subset of the training subsets, wherein the preset brightnesses correspond to the cervical cell sub-image sets; and setting the cervical cell sub-image sets included in each training subset of the training subsets as the training set, so that after a plurality of third cervical cell sub-images are marked on a marking interface, brightness adjustment is performed on the marked third cervical cell sub-images by adopting a plurality of preset brightnesses, so as to obtain the training set, and preparation is performed for recognizing cervical cell images shot under different environmental lights for a subsequent abnormal cell screening model.
302. Constructing a plurality of neural networks to be trained, wherein the plurality of neural networks to be trained correspond to the plurality of neural networks;
the neural networks to be trained respectively run on a plurality of processes, and the processes correspond to the neural networks to be trained.
303. And training the training set based on the plurality of neural networks to be trained so as to obtain the abnormal cell screening model.
In the above technical solution, it can be seen that, by acquiring a training set, the training set includes a plurality of training subsets, where the plurality of training subsets correspond to the plurality of neural networks, each training subset includes a plurality of cervical cell sub-image sets with different brightness degrees, each cervical cell sub-image set includes a plurality of second cervical cell sub-images with different brightness degrees, and each of the plurality of second cervical cell sub-images includes abnormal cells different from each other; constructing a plurality of neural networks to be trained, wherein the plurality of neural networks to be trained correspond to the plurality of neural networks; training the training set based on the plurality of neural networks to be trained to obtain the abnormal cell screening model, so that the abnormal cell screening model can identify cervical cell images shot under different ambient lights, the optimization of the abnormal cell screening model is realized, and the practicability of the abnormal cell screening model is enhanced.
Referring to fig. 4, fig. 4 is a schematic diagram of an abnormal cell screening apparatus according to an embodiment of the present invention. As shown in fig. 4, an abnormal cell screening apparatus 400 according to an embodiment of the present invention may include:
a first acquisition module 401 for acquiring cervical cell images;
optionally, the first acquiring module 401 is configured to acquire the cervical cell image through a scanning device.
The scanning device may be, for example, a scanner. It will be appreciated that a scanner is used to scan the cervical cell layer detected by the liquid-based lamellar cells to obtain a cervical cell image.
Wherein the cervical cell image is an image of a cervical cell layer.
A segmentation module 402, configured to segment the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
optionally, the segmentation module 402 is specifically configured to obtain a gray value corresponding to each pixel point in the cervical cell image; determining a pixel point with the minimum gray value according to the gray value corresponding to each pixel point in the cervical cell image; selecting any one pixel point from the pixel points as a coordinate origin for dividing the cervical cell image; establishing a coordinate system on the cervical cell image based on the origin of coordinates, wherein the coordinate system takes the transverse positive direction of the cervical cell image as an x-axis and the longitudinal positive direction of the cervical cell image as a y-axis; and cutting the cervical cell image from the origin of coordinates to obtain the plurality of cervical cell sub-images.
Optionally, the segmentation module 402 is specifically configured to determine a plurality of gray-scale differences according to the gray-scale value corresponding to each pixel in the cervical cell image, where each gray-scale difference is a difference between each pixel and a corresponding adjacent pixel in the gray-scale value; dividing the gray difference values into a group which falls into the same gray value interval to obtain a plurality of gray groups; determining a plurality of segmentation shapes corresponding to the cervical cell image from the origin of coordinates according to the plurality of gray groups; normalizing the irregular shape in the plurality of segmentation shapes to obtain a plurality of first segmentation shapes with rules; the plurality of first cut-out shapes are set as the plurality of cervical cell sub-images.
Optionally, the segmentation module 402 is specifically configured to obtain a gray value corresponding to each pixel point in the cervical cell image; determining a plurality of groups of pixel points with the same gray value in the cervical cell image according to the gray value corresponding to each pixel point in the cervical cell image, wherein each group of pixel points comprises at least one pixel point, and the gray values corresponding to each pixel point in the at least one pixel point are the same; dividing the cervical cell image according to the plurality of groups of pixel points to obtain a plurality of cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of cervical cell sub-images.
Optionally, the segmentation module 402 is specifically configured to segment the cervical cell image according to the multiple sets of pixel points to obtain multiple second cervical cell sub-images, where the multiple sets of pixel points correspond to the multiple second cervical cell sub-images; determining at least one third cervical cell sub-image having an irregular outline in the plurality of second cervical cell sub-images; performing the following operations for each of the at least one third cervical cell sub-image to obtain the plurality of cervical cell sub-images, including: determining the outer contour size corresponding to the currently processed third cervical cell sub-image; obtaining a template image according to the outer contour size corresponding to the currently processed third cervical cell sub-image so as to obtain the outer contour size corresponding to the template image; and carrying out pixel diffusion on the third cervical cell sub-image which is processed currently according to the outer contour size corresponding to the template image until the outer contour size corresponding to the third cervical cell sub-image which is processed currently is the same as the outer contour size corresponding to the template image, wherein the pixel diffusion is carried out by adopting a first gray value.
An input module 403, configured to input the plurality of cervical cell sub-images into an abnormal cell screening model, so as to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, where each cervical cell sub-image corresponds to a prediction result, and each prediction result is used to indicate an abnormal cell included in each cervical cell sub-image, and the abnormal cell is a cell that generates a lesion or a cancer on the basis of the cervical cell;
wherein, the cervical cells are cells in a normal growth state.
The abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells.
Further, a first neural network is one of the plurality of neural networks, the first neural network being for screening for first abnormal cells included in each of the plurality of cervical cell sub-images, and a second neural network is another of the plurality of neural networks different from the first neural network, the second neural network being for screening for second abnormal cells included in each of the plurality of cervical cell sub-images.
Wherein the first abnormal cell is different from the second abnormal cell, the first abnormal cell is one of the abnormal cells, and the second abnormal cell is another of the abnormal cells.
Alternatively, the abnormal cells include at least one of: squamous carcinoma (squamous cell carcinom, SCC), high-grade squamous epithelial lesions (high-grade squamous epithelial lesion, LSIL), atypical squamous cells that cannot exclude high-grade squamous intraepithelial lesions (atypical squamous cells, cannot exclude high-gradesquamous intraepithelial lesion, ASC-H))), low-grade squamous epithelial lesions (low-grade squamous epithelial lesion, LSIL), atypical squamous cells of undefined significance (atypical squamouscells of unde. Fixed signature, ASC-US), adenocarcinoma (AC), atypical adenocarcinoma (Atypical adenocarcinoma, AGC), and the like.
It will be appreciated that the plurality of neural networks are used to screen each of the plurality of cervical cell sub-images for an abnormal cell. Specifically, a certain neural network of the plurality of neural networks is used to screen squamous cell carcinoma included in each of the plurality of cervical cell sub-images, and another neural network of the plurality of neural networks is used to screen adenocarcinoma included in each of the plurality of cervical cell sub-images.
Optionally, before the plurality of cervical cell sub-images are respectively input into the abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, the abnormal cell screening device further includes a processing module, and the processing module is configured to obtain a training set, where the training set includes a plurality of training subsets, the plurality of training subsets correspond to the plurality of neural networks, each training subset includes a plurality of cervical cell sub-image sets with different brightness degrees, each cervical cell sub-image set includes a plurality of second cervical cell sub-images with one brightness degree, and each second cervical cell sub-image in the plurality of second cervical cell sub-images includes abnormal cells different from each other; constructing a plurality of neural networks to be trained, wherein the plurality of neural networks to be trained correspond to the plurality of neural networks; and training the training set based on the plurality of neural networks to be trained so as to obtain the abnormal cell screening model.
Optionally, the processing module is specifically configured to display the plurality of third cervical cell sub-images on a marker interface; when detecting marking operations for a plurality of positions on the marking interface, marking the plurality of third cervical cell sub-images corresponding to the plurality of positions to obtain a plurality of fourth cervical cell sub-images corresponding to the plurality of third cervical cell sub-images, wherein each fourth cervical cell sub-image is an image obtained by marking each third cervical cell sub-image; processing the fourth cervical cell sub-images with a plurality of preset brightnesses respectively to obtain the cervical cell sub-image sets included in each training subset of the training subsets, wherein the preset brightnesses correspond to the cervical cell sub-image sets; setting the plurality of cervical cell sub-image sets included in each of the plurality of training subsets as the training set.
A selecting module 404, configured to select a plurality of predicted results from the plurality of predicted results according to a preset selection policy as a plurality of first predicted results;
the preset selection strategy is determined according to a prediction selection operation, and the prediction selection operation comprises the following steps: obtaining a prediction probability corresponding to each prediction result in the plurality of prediction results; numbering the prediction probability corresponding to each of the plurality of prediction results according to the order of the prediction probabilities from large to small so as to obtain a plurality of numbers; selecting a part number from the plurality of numbers, wherein the part number is at least one number which is larger than a preset number in the plurality of numbers; and taking the part prediction probability corresponding to the part numbers one by one as the preset selection strategy.
Wherein the preset number is set by an administrator. For example, the number is 10 numbers from 10-1, the preset number is 6, then the partial number is 3 numbers of 7-10.
Further, the selecting a plurality of predicted results from the plurality of predicted results according to a preset selection policy as a plurality of first predicted results includes: and selecting a plurality of prediction results from the plurality of prediction results according to the partial prediction probability as a plurality of first prediction results.
A second obtaining module 405, configured to obtain a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
a display module 406 for displaying the plurality of first cervical cell sub-images on a display interface;
referring to fig. 5, fig. 5 is a schematic server structure diagram of a hardware running environment according to an embodiment of the present application.
The embodiment of the invention provides an electronic device for pushing information, which comprises a processor, a memory, a communication interface and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the processor to execute instructions comprising the steps in any information pushing method. As shown in fig. 5, a server of a hardware running environment according to an embodiment of the present application may include:
a processor 501, such as a CPU.
The memory 502 may alternatively be a high-speed RAM memory or a stable memory, such as a disk memory.
A communication interface 503 for enabling a connected communication between the processor 501 and the memory 502.
Those skilled in the art will appreciate that the structure of the server shown in fig. 5 is not limiting and may include more or fewer components than shown, or may combine certain components, or a different arrangement of components.
As shown in fig. 5, the memory 502 may include an operating system, a network communication module, and a program for data processing. An operating system is a program that manages and controls server hardware and software resources, programs that support personnel management, and other software or program runs. The network communication module is used to enable communication between components within the memory 502 and with other hardware and software within the server.
In the server shown in fig. 5, the processor 501 is configured to execute a program for personnel management stored in the memory 502, and the following steps are implemented:
obtaining a cervical cell image;
segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells;
Selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results;
acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
displaying the plurality of first cervical cell sub-images on a display interface;
the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells. The specific implementation of the server according to the present application may be referred to each embodiment of the above-mentioned abnormal cell screening method, and will not be described herein.
The present application also provides a computer readable storage medium for storing a computer program, the stored computer program being executed by the processor to implement the steps of:
obtaining a cervical cell image;
segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells;
Selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results;
acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
displaying the plurality of first cervical cell sub-images on a display interface;
the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells. The specific implementation of the computer readable storage medium according to the present application can be found in the embodiments of the abnormal cell screening method described above, and will not be described herein.
It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of action combinations, but it should be understood by those skilled in the art that the present invention is not limited by the order of action described, as some steps may be performed in other order or simultaneously in accordance with the present invention. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required for the present invention.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (8)
1. A method for screening abnormal cells, comprising:
obtaining a cervical cell image;
segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
respectively inputting the plurality of cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the plurality of cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with lesions or cancerations on the basis of the cervical cells;
Selecting a plurality of predicted results from the plurality of predicted results according to a preset selection strategy to serve as a plurality of first predicted results;
acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
displaying the plurality of first cervical cell sub-images on a display interface;
the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells;
the segmenting the cervical cell image based on the gray value of the pixel point in the cervical cell image to obtain a plurality of cervical cell sub-images comprises: acquiring a gray value corresponding to each pixel point in the cervical cell image; determining a pixel point with the minimum gray value according to the gray value corresponding to each pixel point in the cervical cell image; selecting any one pixel point from the pixel points as a coordinate origin for dividing the cervical cell image; establishing a coordinate system on the cervical cell image based on the origin of coordinates, wherein the coordinate system takes the transverse positive direction of the cervical cell image as an x-axis and the longitudinal positive direction of the cervical cell image as a y-axis; segmenting the cervical cell image from the origin of coordinates to obtain the plurality of cervical cell sub-images; or alternatively, the first and second heat exchangers may be,
The segmenting the cervical cell image based on the gray value of the pixel point in the cervical cell image to obtain a plurality of cervical cell sub-images comprises: acquiring a gray value corresponding to each pixel point in the cervical cell image; determining a plurality of groups of pixel points with the same gray value in the cervical cell image according to the gray value corresponding to each pixel point in the cervical cell image, wherein each group of pixel points comprises at least one pixel point, and the gray values corresponding to each pixel point in the at least one pixel point are the same; dividing the cervical cell image according to the plurality of groups of pixel points to obtain a plurality of cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of cervical cell sub-images.
2. The method of claim 1, wherein said segmenting the cervical cell image from the origin of coordinates to obtain a plurality of cervical cell sub-images comprises:
determining a plurality of gray difference values according to gray values corresponding to each pixel point in the cervical cell image, wherein each gray difference value is a difference value between each pixel point and a corresponding adjacent pixel point on the gray value;
dividing the gray difference values into a group which falls into the same gray value interval to obtain a plurality of gray groups;
Determining a plurality of segmentation shapes corresponding to the cervical cell image from the origin of coordinates according to the plurality of gray groups;
normalizing the irregular shape in the plurality of segmentation shapes to obtain a plurality of first segmentation shapes with rules;
the plurality of first cut-out shapes are set as the plurality of cervical cell sub-images.
3. The method of claim 1, wherein the segmenting the cervical cell image from the plurality of sets of pixel points to obtain the plurality of cervical cell sub-images comprises:
dividing the cervical cell image according to the plurality of groups of pixel points to obtain a plurality of second cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of second cervical cell sub-images;
determining at least one third cervical cell sub-image having an irregular outline in the plurality of second cervical cell sub-images;
performing the following operations for each of the at least one third cervical cell sub-image to obtain the plurality of cervical cell sub-images, including:
determining the outer contour size corresponding to the currently processed third cervical cell sub-image; obtaining a template image according to the outer contour size corresponding to the currently processed third cervical cell sub-image so as to obtain the outer contour size corresponding to the template image; and carrying out pixel diffusion on the third cervical cell sub-image which is processed currently according to the outer contour size corresponding to the template image until the outer contour size corresponding to the third cervical cell sub-image which is processed currently is the same as the outer contour size corresponding to the template image, wherein the pixel diffusion is carried out by adopting a first gray value.
4. The method of claim 1, wherein prior to said inputting the plurality of cervical cell sub-images into the abnormal cell screening model, respectively, to obtain a plurality of predictions corresponding to the plurality of cervical cell sub-images, the method further comprises:
acquiring a training set, wherein the training set comprises a plurality of training subsets, the plurality of training subsets correspond to the plurality of neural networks, each training subset comprises a plurality of cervical cell sub-image sets with different brightness degrees, each cervical cell sub-image set comprises a plurality of second cervical cell sub-images with different brightness degrees, and each second cervical cell sub-image in the plurality of second cervical cell sub-images comprises abnormal cells different;
constructing a plurality of neural networks to be trained, wherein the plurality of neural networks to be trained correspond to the plurality of neural networks;
and training the training set based on the plurality of neural networks to be trained so as to obtain the abnormal cell screening model.
5. The method of claim 4, wherein the acquiring the training set comprises:
displaying a plurality of third cervical cell sub-images on the label interface;
when detecting marking operations for a plurality of positions on the marking interface, marking the plurality of third cervical cell sub-images corresponding to the plurality of positions to obtain a plurality of fourth cervical cell sub-images corresponding to the plurality of third cervical cell sub-images, wherein each fourth cervical cell sub-image is an image obtained by marking each third cervical cell sub-image;
Processing the fourth cervical cell sub-images with a plurality of preset brightnesses respectively to obtain the cervical cell sub-image sets included in each training subset of the training subsets, wherein the preset brightnesses correspond to the cervical cell sub-image sets;
setting the plurality of cervical cell sub-image sets included in each of the plurality of training subsets as the training set.
6. An abnormal cell screening apparatus comprising:
the first acquisition module is used for acquiring cervical cell images;
the segmentation module is used for segmenting the cervical cell image based on gray values of pixel points in the cervical cell image to obtain a plurality of cervical cell sub-images;
the input module is used for respectively inputting the cervical cell sub-images into an abnormal cell screening model to obtain a plurality of prediction results corresponding to the cervical cell sub-images, wherein each cervical cell sub-image corresponds to one prediction result, each prediction result is used for indicating abnormal cells included in each cervical cell sub-image, and the abnormal cells are cells with pathological changes or cancerations on the basis of the cervical cells;
The selecting module is used for selecting a plurality of predicted results from the plurality of predicted results according to a preset selecting strategy to serve as a plurality of first predicted results;
the second acquisition module is used for acquiring a plurality of first cervical cell sub-images corresponding to the plurality of first prediction results;
a display module for displaying the plurality of first cervical cell sub-images on a display interface;
the abnormal cell screening model comprises a plurality of neural networks, the number of the neural networks is equal to the number of the types of the abnormal cells, and the neural networks are used for screening different abnormal cells;
the segmenting the cervical cell image based on the gray value of the pixel point in the cervical cell image to obtain a plurality of cervical cell sub-images comprises: acquiring a gray value corresponding to each pixel point in the cervical cell image; determining a pixel point with the minimum gray value according to the gray value corresponding to each pixel point in the cervical cell image; selecting any one pixel point from the pixel points as a coordinate origin for dividing the cervical cell image; establishing a coordinate system on the cervical cell image based on the origin of coordinates, wherein the coordinate system takes the transverse positive direction of the cervical cell image as an x-axis and the longitudinal positive direction of the cervical cell image as a y-axis; segmenting the cervical cell image from the origin of coordinates to obtain the plurality of cervical cell sub-images; or alternatively, the first and second heat exchangers may be,
The segmenting the cervical cell image based on the gray value of the pixel point in the cervical cell image to obtain a plurality of cervical cell sub-images comprises: acquiring a gray value corresponding to each pixel point in the cervical cell image; determining a plurality of groups of pixel points with the same gray value in the cervical cell image according to the gray value corresponding to each pixel point in the cervical cell image, wherein each group of pixel points comprises at least one pixel point, and the gray values corresponding to each pixel point in the at least one pixel point are the same; dividing the cervical cell image according to the plurality of groups of pixel points to obtain a plurality of cervical cell sub-images, wherein the plurality of groups of pixel points correspond to the plurality of cervical cell sub-images.
7. An electronic device for abnormal cell screening, comprising a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and are generated for execution by the processor to perform the instructions of the steps of the method of any of claims 1-5.
8. A computer readable storage medium for storing a computer program for execution by a processor to implement the method of any one of claims 1-5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911040955.9A CN110992303B (en) | 2019-10-29 | 2019-10-29 | Abnormal cell screening method and device, electronic equipment and storage medium |
| PCT/CN2020/093581 WO2021082434A1 (en) | 2019-10-29 | 2020-05-30 | Abnormal cell screening method and apparatus, electronic device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911040955.9A CN110992303B (en) | 2019-10-29 | 2019-10-29 | Abnormal cell screening method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110992303A CN110992303A (en) | 2020-04-10 |
| CN110992303B true CN110992303B (en) | 2023-12-22 |
Family
ID=70082495
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911040955.9A Active CN110992303B (en) | 2019-10-29 | 2019-10-29 | Abnormal cell screening method and device, electronic equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110992303B (en) |
| WO (1) | WO2021082434A1 (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110415212A (en) * | 2019-06-18 | 2019-11-05 | 平安科技（深圳）有限公司 | Abnormal cell detection method, device and computer readable storage medium |
| CN110992303B (en) * | 2019-10-29 | 2023-12-22 | 平安科技（深圳）有限公司 | Abnormal cell screening method and device, electronic equipment and storage medium |
| CN111951271B (en) * | 2020-06-30 | 2023-12-15 | 杭州依图医疗技术有限公司 | Method and device for identifying cancer cells in pathological image |
| CN111967465A (en) * | 2020-07-07 | 2020-11-20 | 广州金域医学检验中心有限公司 | Method, system, computer device and storage medium for evaluating tumor cell content |
| CN112037862B (en) * | 2020-08-26 | 2021-11-30 | 深圳太力生物技术有限责任公司 | Cell screening method and device based on convolutional neural network |
| CN112184733B (en) * | 2020-09-30 | 2024-07-23 | 平安科技（深圳）有限公司 | Cervical abnormal cell detection device and method |
| CN113111926A (en) * | 2021-03-31 | 2021-07-13 | 南京华晟医学检验实验室有限公司 | Abnormal cervical blood cell screening method based on TCT (TCT) slide |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016189469A1 (en) * | 2015-05-25 | 2016-12-01 | Adarsh Natarajan | A method for medical screening and a system therefor |
| WO2018008593A1 (en) * | 2016-07-04 | 2018-01-11 | 日本電気株式会社 | Image diagnosis learning device, image diagnosis device, image diagnosis method, and recording medium for storing program |
| CN107742299A (en) * | 2017-11-28 | 2018-02-27 | 中国联合网络通信集团有限公司 | A kind of image partition method and device |
| CN109872306A (en) * | 2019-01-28 | 2019-06-11 | 腾讯科技（深圳）有限公司 | Medical image cutting method, device and storage medium |
| US10354122B1 (en) * | 2018-03-02 | 2019-07-16 | Hong Kong Applied Science and Technology Research Institute Company Limited | Using masks to improve classification performance of convolutional neural networks with applications to cancer-cell screening |
| CN110120040A (en) * | 2019-05-13 | 2019-08-13 | 广州锟元方青医疗科技有限公司 | Sectioning image processing method, device, computer equipment and storage medium |
| CN110211108A (en) * | 2019-05-29 | 2019-09-06 | 武汉兰丁医学高科技有限公司 | A kind of novel abnormal cervical cells automatic identifying method based on Feulgen colouring method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008005426A2 (en) * | 2006-06-30 | 2008-01-10 | University Of South Florida | Computer-aided pathological diagnosis system |
| CN105095865A (en) * | 2015-07-17 | 2015-11-25 | 广西师范大学 | Directed-weighted-complex-network-based cervical cell recognition method and a cervical cell recognition apparatus |
| CN110689518B (en) * | 2019-08-15 | 2024-07-02 | 平安科技（深圳）有限公司 | Cervical cell image screening method, cervical cell image screening device, computer equipment and storage medium |
| CN110992303B (en) * | 2019-10-29 | 2023-12-22 | 平安科技（深圳）有限公司 | Abnormal cell screening method and device, electronic equipment and storage medium |
-
2019
- 2019-10-29 CN CN201911040955.9A patent/CN110992303B/en active Active
-
2020
- 2020-05-30 WO PCT/CN2020/093581 patent/WO2021082434A1/en active Application Filing
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016189469A1 (en) * | 2015-05-25 | 2016-12-01 | Adarsh Natarajan | A method for medical screening and a system therefor |
| WO2018008593A1 (en) * | 2016-07-04 | 2018-01-11 | 日本電気株式会社 | Image diagnosis learning device, image diagnosis device, image diagnosis method, and recording medium for storing program |
| CN107742299A (en) * | 2017-11-28 | 2018-02-27 | 中国联合网络通信集团有限公司 | A kind of image partition method and device |
| US10354122B1 (en) * | 2018-03-02 | 2019-07-16 | Hong Kong Applied Science and Technology Research Institute Company Limited | Using masks to improve classification performance of convolutional neural networks with applications to cancer-cell screening |
| CN109872306A (en) * | 2019-01-28 | 2019-06-11 | 腾讯科技（深圳）有限公司 | Medical image cutting method, device and storage medium |
| CN110120040A (en) * | 2019-05-13 | 2019-08-13 | 广州锟元方青医疗科技有限公司 | Sectioning image processing method, device, computer equipment and storage medium |
| CN110211108A (en) * | 2019-05-29 | 2019-09-06 | 武汉兰丁医学高科技有限公司 | A kind of novel abnormal cervical cells automatic identifying method based on Feulgen colouring method |
Non-Patent Citations (1)
| Title |
|---|
| 基于深度卷积神经网络的宫颈细胞病理智能辅助诊断方法;廖欣 等;液晶与显示;第33卷(第06期);第528-537页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021082434A1 (en) | 2021-05-06 |
| CN110992303A (en) | 2020-04-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110992303B (en) | Abnormal cell screening method and device, electronic equipment and storage medium | |
| CN111652845B (en) | Automatic labeling method and device for abnormal cells, electronic equipment and storage medium | |
| CN110060237B (en) | Fault detection method, device, equipment and system | |
| CN107067402B (en) | Medical image processing apparatus and breast image processing method thereof | |
| CN110866529B (en) | Character recognition method, device, electronic equipment and storage medium | |
| CN113283446B (en) | Method and device for identifying object in image, electronic equipment and storage medium | |
| CN109919903B (en) | Spine detection positioning marking method and system and electronic equipment | |
| CN105719243B (en) | Image processing apparatus and method | |
| CN112508850A (en) | Deep learning-based method for detecting malignant area of thyroid cell pathological section | |
| CN112699775A (en) | Certificate identification method, device and equipment based on deep learning and storage medium | |
| CN113033543B (en) | Curve text recognition method, device, equipment and medium | |
| US11600088B2 (en) | Utilizing machine learning and image filtering techniques to detect and analyze handwritten text | |
| CN112785591B (en) | Method and device for detecting and segmenting rib fracture in CT image | |
| TWI587844B (en) | Medical image processing apparatus and breast image processing method thereof | |
| CN113487621A (en) | Medical image grading method and device, electronic equipment and readable storage medium | |
| CN112132781A (en) | Texture defect detection method and device | |
| Shanker et al. | Brain tumor segmentation of normal and lesion tissues using hybrid clustering and hierarchical centroid shape descriptor | |
| CN115908363A (en) | Tumor cell counting method, device, equipment and storage medium | |
| US11176419B1 (en) | Method for labeling image | |
| CN117372415A (en) | Laryngoscope image recognition method, device, computer equipment and storage medium | |
| CN112528903A (en) | Face image acquisition method and device, electronic equipment and medium | |
| CN117522773A (en) | Product detection method, electronic device and storage medium | |
| CN112614138A (en) | Image processing apparatus, image processing system, storage medium, and image processing method | |
| CN115294426B (en) | Method, device and equipment for tracking interventional medical equipment and storage medium | |
| CN116825269A (en) | Physical examination report processing method and device, electronic equipment and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40017509 Country of ref document: HK |
|
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |