CN115880233A - Cytopathology image evaluation method and device - Google Patents
Cytopathology image evaluation method and device Download PDFInfo
- Publication number
- CN115880233A CN115880233A CN202211502167.9A CN202211502167A CN115880233A CN 115880233 A CN115880233 A CN 115880233A CN 202211502167 A CN202211502167 A CN 202211502167A CN 115880233 A CN115880233 A CN 115880233A
- Authority
- CN
- China
- Prior art keywords
- image
- result
- evaluation
- cytopathology
- scanning
- 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.)
- Pending
Links
Images
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention provides a cytopathology image evaluation method and a cytopathology image evaluation device, wherein a cytopathology mounting is obtained, the cytopathology mounting is scanned, a digital scanning image is obtained, the digital scanning image is used as the input of a digital pathology evaluation model, the evaluation result of the scanning quality of the digital pathology image is output, and the evaluation result comprises the following steps: the sample wafer result, the result to be reproduced and the result to be rescanned can screen out cytopathology images with unqualified wafer production quality and unqualified scanning quality in time, reproduce cell mounting with unqualified quality in time, rescan the cytopathology mounting with unqualified quality in time, correct the cytopathology images in time, reduce report generation time and improve timeliness of evaluating the cytopathology images. In addition, through deep learning of a computer, the cytopathology image can be objectively evaluated through machine assistance, the subjectivity of evaluation is reduced, and the consistency of evaluation is improved.
Description
Technical Field
The invention relates to the technical field of cell engineering artificial intelligence, in particular to a cell pathology image evaluation method and device.
Background
Cytopathology is an important screening means for neoplastic diseases in clinical diagnosis, and is an important means for observing tumorigenesis and early prevention and treatment. Among them, cervical cancer is one of the high-incidence malignant tumors in the world, and early discovery, early diagnosis and early treatment are important means for preventing and improving the current state of cervical cancer.
At present, pathological diagnosis mainly depends on cell case images, all processes such as control sheet, dyeing, mounting, scanning and the like need to be performed, defects in any link affect diagnosis work, and the processing timeliness of the processes with the defects is poor; in addition, the existing cytopathology image evaluation has strong subjective dependence, poor consistency of evaluation results and poor repeatability, a certain time is required from the film production to the film reading, and a certain time is also required from the film reading to the feedback, so that the poor image quality cannot be fed back in time during the film production, and the diagnosis quality of computer-aided screening work is reduced.
Therefore, finding an adaptive cytopathology image evaluation method is a technical problem to be solved urgently by those skilled in the art.
Disclosure of Invention
In view of the above-mentioned deficiencies of the prior art, the present invention aims to provide a cytopathology evaluation method and apparatus, which can solve the problems of poor timeliness, strong subjectivity, poor consistency and poor repeatability in evaluating cytopathology images.
In a first aspect, the present invention provides a cytopathology image evaluation method, including the steps of: obtaining a cytopathology mounting, scanning the cytopathology mounting and obtaining a digital scanning image; taking the digital scanning image as an input of a digital pathological image evaluation model, and outputting an evaluation result of the scanning quality of the digital pathological image, wherein the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
Further, the step of obtaining a cytopathology envelope comprises: acquiring the visual field of the cell pathological section by adjusting the coordinate parameters of the equipment; analyzing the amount of cells in the field of view; if the cell amount is larger than a preset cell amount threshold value, analyzing the staining quality of the cells in the visual field; if the dyeing quality is larger than a preset RGB value, analyzing the number of bubbles in the visual field; and if the number of the bubbles is not more than a preset bubble threshold value, obtaining the cytopathology mounting plate.
Further, the step of scanning the cytopathological mount and acquiring a digitally scanned image comprises: scanning the cytopathology mounting sheet to obtain a first image, analyzing the scanning definition of the first image, analyzing the scanning color of the first image if the scanning definition is greater than a preset definition threshold, and taking the first image as a digital scanning image if the scanning color is greater than a preset color threshold.
Further, if the evaluation result is a positive sample result, outputting the positive sample result; if the evaluation result is the result to be reproduced, acquiring a second cytopathology mounting, scanning the second cytopathology mounting and acquiring a second image, taking the second image as the input of the digital pathology image evaluation model, and outputting the evaluation result of the second image; and if the evaluation result is the result to be rescanned, rescanning the cytopathology mounting and acquiring a third image, taking the third image as the input of the digital pathology image evaluation model, and outputting the evaluation result of the third image.
Further, if the evaluation result is a positive sample result, analyzing the positive sample result as a cell quality evaluation result and an image quality evaluation result, and inputting the cell quality evaluation result and the image quality evaluation result as a multi-dimensional evaluation model; if the evaluation result is the result to be reproduced, acquiring a second cytopathology mounting, scanning the second cytopathology mounting and acquiring a second image, taking the second image as the input of the digital pathology image evaluation model, outputting the second evaluation result of the second image, analyzing the second evaluation result as a cell quality evaluation correction result, and taking the cell quality evaluation correction result as the input of the multi-dimensional evaluation model; and if the evaluation result is the result to be rescanned, rescanning the cytopathology mounting and acquiring a third image, taking the third image as the input of the digital pathology image evaluation model, outputting the third evaluation result of the third image, analyzing the third evaluation result as an image quality evaluation correction result, and taking the image evaluation correction result as the input of the multi-dimensional evaluation model.
Further, the step of inputting the cell quality evaluation correction result as a multidimensional evaluation model includes: analyzing the cell quality evaluation correction result, and outputting a pathological evaluation result of the multidimensional evaluation model according to the cell quality evaluation correction result, wherein the pathological evaluation result comprises a positive result and a negative result; the step of inputting the image evaluation correction result as the input of the multidimensional evaluation model comprises the following steps: analyzing the image evaluation and correction result, and outputting the image evaluation result of the multi-dimensional evaluation model according to the image evaluation and correction result, wherein the image evaluation result comprises the result of the image to be retrieved.
Further, the scanning the cytopathology mounting obtains the first image, analyzes the scanning definition of the first image, analyzes the scanning color of the first image if the scanning definition is greater than a preset definition threshold, and takes the first image as a digital scanning image if the scanning color is greater than a preset color threshold, including: scanning the cytopathology mounting to obtain a first image, dividing the first image into a plurality of first sub-images, and calculating the sub-definition of each first sub-image; if the sub-definition of the first sub-image is greater than a preset sub-definition threshold, judging that the first sub-image is a clear first sub-image, and if the number of the first sub-images in the first image, of which the sub-definition is greater than the preset sub-definition threshold, reaches a preset number, then the scanning definition of the first image is greater than the preset definition threshold, and then the scanning color of the first image is analyzed; comparing the RGB parameters of the first image with preset RGB parameters, outputting an RGB parameter comparison result, and if the RGB parameter comparison result exceeds an RGB preset standard deviation range, taking the first image as a digital scanning image if the scanning color is greater than a preset color threshold value.
Further, the step of obtaining a cytopathology mounting piece if the number of bubbles is not greater than a preset bubble threshold value includes: calculating the sub-definition difference value of the adjacent first sub-images in the first image, if the sub-definition difference value is larger than a preset sub-definition difference value, screening out a target area with the sub-definition difference value larger than the preset sub-definition difference value, calculating the number of bubbles, and if the number of bubbles is not larger than a preset bubble threshold value, acquiring the cytopathology mounting.
Further, the step of inputting the digital scanning image as an input of the digital pathology image evaluation model includes: acquiring qualified digital pathological images, and establishing a training data set of a proof sample; acquiring unqualified digital pathological images, and establishing a loading piece training data set; inputting a positive sample training data set and a negative sample training data set into a convolutional neural network model for training; and acquiring the trained convolutional neural network model as a digital pathological image evaluation model.
In a second aspect, the present invention provides a cytopathology image evaluation device including: the acquisition module is used for acquiring a cytopathology mounting, scanning the cytopathology mounting and acquiring a digital scanning image; an evaluation model module, configured to take the digital scan image as an input of the digital pathology image evaluation model, and output an evaluation result of the digital pathology image scan quality, where the evaluation result includes: a positive sample result, a result to be reproduced, and a result to be rescanned.
According to the scheme provided by the invention, a cytopathology mounting sheet is obtained, the cytopathology mounting sheet is scanned, a digital scanning image is obtained, the digital scanning image is used as the input of a digital pathology evaluation model, and the evaluation result of the scanning quality of the digital pathology image is output, wherein the evaluation result comprises the following steps: the sample wafer result, the result to be reproduced and the result to be rescanned can screen out cytopathology images with unqualified wafer production quality and unqualified scanning quality in time, reproduce cell mounting with unqualified quality in time, rescan the cytopathology mounting with unqualified quality in time, correct the cytopathology images in time, reduce report generation time and improve timeliness of evaluating the cytopathology images. In addition, through deep learning of a computer, the cytopathology image can be objectively evaluated through machine assistance, the subjectivity of evaluation is reduced, and the consistency of evaluation is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a cytopathology image evaluation method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a cytopathology envelope in accordance with one embodiment of the present invention;
FIG. 3 is a schematic diagram of a qualifying cytopathology image in accordance with an embodiment of the invention;
FIG. 4 is a schematic diagram of the achievement of a desired cell mass in one embodiment of the present invention;
FIG. 5 is a diagram of a sample of the cells meeting the standards in one embodiment of the present invention;
FIG. 6 is a schematic diagram showing the failure to achieve the desired cell mass according to one embodiment of the present invention;
FIG. 7 is a diagram of a sample of an embodiment of the present invention showing an substandard cell count;
FIG. 8 is a schematic view of a wafer with bubbles in accordance with an embodiment of the present invention;
FIG. 9 is a pictorial view of an embodiment of the present invention with the bubble in the mounting plate;
FIG. 10 is a schematic flow chart of a method for evaluating a cytopathology image according to an embodiment of the present invention;
FIG. 11 is a schematic flow chart of a method for evaluating a cytopathology image according to an embodiment of the present invention;
FIG. 12 is a flow chart of a cytopathology image evaluating apparatus according to an embodiment of the present invention;
FIG. 13 is a block diagram of the internal structure of a computer device in one embodiment of the invention.
Reference numerals:
1. squamous epithelial cells.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or to implicitly indicate the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present application, "plurality" or "a plurality" means two or more unless specifically limited otherwise.
It should be understood that the structures, ratios, sizes, and the like shown in the drawings are only used for matching the disclosure of the specification, so as to be understood and read by those skilled in the art, and are not used to limit the practical limit conditions of the present application, so that the modifications of the structures, the changes of the ratio relationships, or the adjustment of the sizes, do not have the technical essence, and the modifications, the changes of the ratio relationships, or the adjustment of the sizes, are all within the scope of the technical contents disclosed in the present application without affecting the efficacy and the achievable purpose of the present application.
Cytopathology is an important screening means for neoplastic diseases in clinical diagnosis, and is an important means for observing tumorigenesis and early prevention and treatment. Among them, cervical cancer is one of the high-incidence malignant tumors in the world, and early discovery, early diagnosis and early treatment are important means for preventing and improving the current state of cervical cancer. At present, the screening of cervical cell pathology by machine learning and deep learning using computer technology is also in the process of transition from research to clinical practice. When the digitized cytopathology image is analyzed by a computer, the cytopathology slice needs to be digitally scanned. However, both the quality of the production and the quality of the scan are important factors affecting the digital cytopathology images. The quality of digital cytopathology images is improved, and the method becomes an important guarantee for improving the cervical cancer screening quality, so that the screening accuracy is improved, and the missed diagnosis rate is reduced. How to use computer technology to perform automatic quality control and evaluation in the digital scanning engineering becomes a current ongoing challenge.
Example one
Referring to fig. 1, the present embodiment provides a cytopathology image evaluation method, including the following steps: obtaining a cytopathology mounting, scanning the cytopathology mounting and obtaining a digital scanning image; taking the digital scanning image as an input of a digital pathological image evaluation model, and outputting an evaluation result of the scanning quality of the digital pathological image, wherein the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
Specifically, a cytopathology tissue section is obtained, the cytopathology tissue section is stained, the stained cytopathology tissue section is mounted, a cytopathology mounting sheet is obtained by referring to the mounted cytopathology tissue section shown in fig. 2, and the cytopathology mounting sheet is scanned in a digital pathology scanner in a conventional manner to generate a digital pathology image; taking the digital scanning image as an input of a digital pathological image evaluation model, performing deep learning auxiliary evaluation by using the digital pathological image evaluation model, and outputting an evaluation result of the scanning quality of the digital pathological image, wherein the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
Referring to fig. 3, a schematic diagram of a qualified digital scan image is shown.
The cytopathology images are divided into cytopathology and histopathology, the cells of the cytopathology are scattered, and the difference between the number of the cells on a section and the cell density is large; histopathology, the biopsy samples are the whole distribution of blocks and pieces, the area of the selected pathological section of the biopsy sample basically reflects the cell amount contained in the specimen.
Taking cervical cytology as an example, respectively calculating the number of squamous epithelial cells 1 and glandular epithelial cells;
taking thyroid cytology as an example, the number of thyroid follicular epithelial cells was calculated.
The problem to be solved by this embodiment is that the existing evaluation method is poor in timeliness, a certain time is required from the film production to the film reading, and a certain time is also required from the film reading to the feedback, so that the poor image quality cannot be fed back in time during the film production, and the diagnosis quality of the computer-aided screening work is reduced. The cytopathology image evaluation process comprises the following steps: the method comprises a flaking process and a scanning process, wherein the flaking process needs to evaluate the cell amount, the dyeing quality and the mounting quality in sequence; the scanning process requires sequential evaluation of the scan sharpness and the scan color difference. Each evaluation step is substandard and requires re-production or re-scanning so that the final cytopathology image is a satisfactory high quality image.
According to the scheme provided by the embodiment, by acquiring a cytopathology mounting, scanning the cytopathology mounting and acquiring a digital scanning image, taking the digital scanning image as an input of a digital pathology evaluation model, and outputting an evaluation result of the scanning quality of the digital pathology image, the evaluation result comprises: the sample film result, the result to be reproduced and the result to be scanned are selected, the cytopathology images with the film making quality not up to the standard and the scanning quality not up to the standard can be screened out in time, the cell mounting films with the quality not up to the standard are reproduced in time, the cytopathology mounting films with the quality not up to the standard are rescanned in time, correction work is carried out in time, report generation time is shortened, and timeliness of evaluating the cytopathology images is improved. In addition, through computer deep learning, the cytopathology image can be objectively evaluated by machine assistance, so that the subjectivity of evaluation is reduced, and the consistency of evaluation is improved.
Example two
The present example provides a further approach.
This embodiment provides a further aspect, wherein the step of obtaining a cytopathology envelope comprises: acquiring the visual field of the cell pathological section by adjusting the coordinate parameters of the equipment; analyzing the amount of cells in the field of view; if the cell amount is larger than a preset cell amount threshold value, analyzing the staining quality of the cells in the visual field; if the dyeing quality is larger than a preset RGB value, analyzing the number of bubbles in the visual field; and if the number of the bubbles is not more than the preset bubble threshold value, obtaining the cytopathology mounting sheet.
Specifically, taking a cervical cytopathology image of a cytopathology as an example, although the total area of the cytopathology section is consistent, the number and density of cells are not consistent, therefore, 20 fields of the cytopathology section are randomly screened, the fields are sub-image blocks of the segmented cytopathology section image, the 20 fields are selected from a central area and a peripheral area of the cytopathology section, and the number of the squamous epithelial cells 1 and the glandular epithelial cells in the fields are respectively calculated by taking the cervical cytology as an example; if the cell amount of the epithelial cells in the cervical cytopathology image is greater than the preset cell amount threshold, for example, the preset cell amount threshold is 5000, the cell amount in the cervical cytopathology image is judged to be qualified, and the model diagram of the digital pathology image shown in fig. 4 and the physical diagram of the digital pathology image shown in fig. 5 are referred to, such as the squamous epithelial cells 1 shown in fig. 5, and the cytopathology images with the normal cell amount of the cell amount greater than 5000 are shown in fig. 4 and 5.
Referring to the schematic diagram of the digital pathology image shown in fig. 6, referring to the physical diagram of the digital pathology image shown in fig. 7, such as the squamous epithelial cells 1 shown in fig. 7, fig. 6 and 7 show the visual field diagram of the cytopathology image with too small cell amount, the squamous epithelial cells 1 in the cytopathology image are too small, the cell distribution is sparse, the cell amount of the whole cytopathology image with too small cell amount is less than 5000,
after the cell amount in the cytopathology image is judged to be qualified, the staining quality of the cells in the visual field is analyzed, of course, the cell amount in the visual field and the cell staining quality in the visual field can be analyzed in parallel, and the sequence of the analysis and the analysis can be switched and also can be processed in parallel. Analyzing the staining quality of the cells in the visual field, wherein the difference of the staining quality is mainly shown on cells with tangible components in the cytopathology image, and the cells with tangible components comprise: obtaining a scanning image of a stained cytopathology section by squamous epithelial cells 1, glandular epithelial cells, basal cells, neutrophils and histiocytes, judging whether the staining quality of the digital scanning image reaches a preset RGB value or not by a digital pathology image evaluation model, and comparing the values: calculating RGB parameters of the digital scanning image, and comparing the RGB parameters with preset RGB values. The method for acquiring the preset RGB value comprises the following steps: selecting 1000 positive samples with qualified dyeing, and presetting a preset RGB value according to the average RGB value of the positive samples. In a further embodiment of this embodiment, if the staining quality is within a predetermined RGB difference range, such as ± 2s (standard deviation), the cytopathology slice corresponding to the digital scanned image is judged to be a slice with a staining quality meeting the standard, and the RGB value of the digital scanned image is also compared with the predetermined RGB value.
In addition, when the seal sheet contains bubbles, the difference in the resolution between the inside and outside of the bubbles appearing in the image of the cell case is large, and the step of analyzing the number of bubbles in the visual field includes: dividing the cytopathology image into a plurality of sub-images, calculating the sub-definition of each sub-image as z, judging the z to be a value between 0 and 1, if the z of the sub-image is greater than 0.9, judging the sub-image to be a clear sub-image, meanwhile, calculating the definition difference of the adjacent sub-image, if the definition difference is greater than the preset definition difference, judging that the area of the adjacent sub-image has bubbles, for example, the definition of a third sub-image is 0.1, the definition of a fourth sub-image is 0.9, the definition difference of the third sub-image and the fourth sub-image is 0.8, and the preset definition difference is 0.01, wherein the definition difference is greater than the preset definition difference; and judging that the area of the adjacent sub-image has no bubble if the definition difference value is smaller than the preset definition difference value. Referring to FIG. 8, a schematic view of the envelope with bubbles is shown, and to FIG. 9, a schematic view of the envelope with bubbles is shown.
The problem to be solved by this embodiment is that staining and then mounting the section is an important process in obtaining a qualified cytopathology image, the cell amount does not reach the standard and/or the staining quality is not good and/or the mounting has bubbles and the like, which are unqualified cell case images, if the unqualified specific reasons cannot be screened in time, more labor cost and time cost will be consumed, and at present, for the digital cytopathology image with unqualified quality, there is no relevant objectivity improvement suggestion.
According to the scheme provided by the embodiment, the cell pathological section corresponding to the cell pathological image with the unqualified cell quantity can be screened in time, and the cell pathological section needs to be produced again; the cell pathological section corresponding to the cell pathological image with poor dyeing quality can be screened in time, the cell pathological section needs to be dyed again, the cell pathological section corresponding to the cell pathological image with air bubbles in the mounting piece can be screened in time, and the cell pathological section needs to be mounted again. The scheme can provide specific improvement suggestions, and when unqualified cytopathology images are obtained, the initial steps are not needed, so that the timeliness of obtaining diagnosis reports is greatly improved.
EXAMPLE III
This embodiment provides a further aspect, the step of scanning the cytopathological mount and acquiring a digitally scanned image, comprising: scanning the cytopathology mounting sheet to obtain a first image, analyzing the scanning definition of the first image, analyzing the scanning color of the first image if the scanning definition is greater than a preset definition threshold, and taking the first image as a digital scanning image if the scanning color is greater than a preset color threshold.
Specifically, the step of analyzing the scanning resolution of the first image includes: the method comprises the steps of dividing a first image into a plurality of first sub-images, judging and reading a definition value z of each first sub-image to be a value between 0 and 1, judging and reading the first sub-image to be a clearly scanned digital scanning image if the definition value of the first sub-image is larger than 0.9, and judging and reading the first image to be a clearly scanned digital scanning image if the definition values of 90% of the first sub-images in the first image are larger than 0.9. Under the condition of scanning local defocus, the definition values of the adjacent first sub-images in the defocus area are gradually increased from the defocus center to the edge area, for example, the definition values from the defocus center to the edge area are 0.2, 0.3, … …, 0.8 and 0.9 in sequence, the definition of the defocus center is low, and the image becomes clearer as the image gets closer to the edge. The step of resolving the scanned color of the first image comprises: obtaining 1000 samples with qualified dyeing quality, scanning the 1000 samples, respectively obtaining RGB values of the 1000 samples, calculating an average RGB value according to the RGB values of the 1000 samples to preset a preset RGB value, comparing the scanning color of the first image with the preset RGB value, and judging that the scanning color of the first image is qualified if the comparison result is within the range of +/-2 s.
In a further scheme, in the cytopathology image with the standard cell amount, the standard dyeing quality, no bubble in mounting, standard definition and poor scanning color, although the average color difference is formed between the evaluation dyeing quality and the evaluation scanning color, the dyeing quality of the tangible cells in the image is mainly evaluated when the dyeing quality is evaluated; poor scanning color is mainly fed back to a blank area of the digital scanning image, stable white balance is established aiming at the blank area, and excessive deviation on tangible cell staining is not influenced.
The digital pathological image evaluation model is evaluated into a qualified digital pathological image, and then the qualified digital pathological image is subjected to preliminary auxiliary diagnosis and analysis through computer-aided interpretation, the analysis result and the digital pathological image are directly fed back to a pathological doctor for reading, the pathological doctor evaluates the digital pathological image, most of the digital pathological image is subjected to quality evaluation of the model, if special conditions exist, the pathological doctor feeds back the digital pathological image as not reaching the standard, the special digital pathological image is classified into a negative sample training set again, and the algorithm of the model is continuously improved. If the model evaluation result is consistent with the pathological doctor evaluation result, the doctor issues a report and issues a positive result or a negative result. As shown in fig. 10.
Example four
The embodiment provides a further scheme, and if the evaluation result is a positive sample result, the positive sample result is output; if the evaluation result is the result to be reproduced, acquiring a second cytopathology mounting, scanning the second cytopathology mounting and acquiring a second image, taking the second image as the input of the digital pathology image evaluation model, and outputting the evaluation result of the second image; and if the evaluation result is the result to be rescanned, rescanning the cytopathology mounting and acquiring a third image, taking the third image as the input of the digital pathology image evaluation model, and outputting the evaluation result of the third image.
Specifically, the cytopathology images corresponding to the to-be-reproduced result comprise cytopathology images with unqualified cell amount, cytopathology images with unqualified dyeing quality and cytopathology images with bubbles in mounting, proper sections are selected for re-production when the cell amount does not reach the standard, the cytopathology sections are re-dyed when the dyeing quality does not reach the standard, and the mounting is re-mounted when bubbles exist. The cytopathology images corresponding to the results to be rescanned comprise cytopathology images with low scanning definition, scanning defocused cytopathology images and cytopathology images with poor scanning color, the slide needs to be cleaned and rescanned under the conditions of low scanning definition and defocused scanning, and the scanning parameters need to be rescanned when the scanning color is poor.
EXAMPLE five
In this embodiment, a further scheme is provided, where if the evaluation result is a positive sample result, the positive sample result is analyzed as a cell quality evaluation result and an image quality evaluation result, and the cell quality evaluation result and the image quality evaluation result are used as inputs of the multidimensional evaluation model; if the evaluation result is the result to be reproduced, acquiring a second cytopathology mounting, scanning the second cytopathology mounting and acquiring a second image, taking the second image as the input of the digital pathology image evaluation model, outputting the second evaluation result of the second image, analyzing the second evaluation result as a cell quality evaluation correction result, and taking the cell quality evaluation correction result as the input of the multi-dimensional evaluation model; and if the evaluation result is the result to be rescanned, rescanning the cytopathology mounting and acquiring a third image, taking the third image as the input of the digital pathology image evaluation model, outputting the third evaluation result of the third image, analyzing the third evaluation result as an image quality evaluation correction result, and taking the image evaluation correction result as the input of the multidimensional evaluation model. As shown in fig. 11.
The problem to be solved by this embodiment is that, specifically, the digital scanning image that is evaluated by the digital pathological image evaluation model to be up to standard in cell amount, up to standard in dyeing quality, no bubble in mounting, up to standard in scanning definition, and up to standard in scanning color is output as a positive sample result, and under special circumstances, the total cell amount is up to standard, but the target cell amount to be tracked is not up to standard, the visual field needs to be traversed, and the cell quality evaluation correction result and the image quality evaluation correction result are obtained. In order to improve the accuracy of diagnosis, and more carefully search for digital scanning images with substandard quality control, and provide improvement suggestions in time.
The cytopathology image evaluation process comprises the following steps: the method comprises a flaking process and a scanning process, wherein the flaking process needs to evaluate the cell amount, the dyeing quality and the mounting quality in sequence; the scanning process requires sequential evaluation of the scan sharpness and the scan color difference. Each evaluation step is substandard and requires re-production or re-scanning so that the final cytopathology image is a satisfactory high quality image.
According to the scheme provided by the embodiment, the diagnosis accuracy and the diagnosis efficiency can be further improved.
EXAMPLE six
This embodiment provides a further solution, wherein the step of using the cell quality evaluation correction result as an input of the multidimensional evaluation model comprises: analyzing the cell quality evaluation correction result, and outputting a pathological evaluation result of the multidimensional evaluation model according to the cell quality evaluation correction result, wherein the pathological evaluation result comprises a positive result and a negative result; the step of using the image evaluation correction result as the input of the multidimensional evaluation model comprises the following steps: analyzing the image evaluation and correction result, and outputting the image evaluation result of the multi-dimensional evaluation model according to the image evaluation and correction result, wherein the image evaluation result comprises the result of the image to be retrieved.
EXAMPLE seven
This embodiment provides a further scheme, the scanning the cytopathology mounting obtains first image, the scanning definition of analysis first image, if the scanning definition is greater than preset definition threshold, then the scanning colour of analysis first image, if the scanning colour is greater than preset colour threshold, then with the step of first image as digital scanning image, include: scanning the cytopathology mounting to obtain a first image, dividing the first image into a plurality of first sub-images, and calculating the sub-definition of each first sub-image; if the sub-definition of the first sub-image is greater than a preset sub-definition threshold, judging that the first sub-image is a clear first sub-image, and if the number of the first sub-images in the first image, of which the sub-definition is greater than the preset sub-definition threshold, reaches a preset number, then the scanning definition of the first image is greater than the preset definition threshold, and then the scanning color of the first image is analyzed; and comparing the RGB parameters of the first image with preset RGB parameters, outputting an RGB parameter comparison result, and if the RGB parameter comparison result exceeds the RGB preset standard difference range, taking the first image as a digital scanning image if the scanning color is larger than a preset color threshold value.
Example eight
Further, the step of obtaining a cytopathology mounting piece if the number of bubbles is not greater than a preset bubble threshold value includes: calculating the sub-definition difference value of the adjacent first sub-images in the first image, if the sub-definition difference value is larger than a preset sub-definition difference value, screening out a target area with the sub-definition difference value larger than the preset sub-definition difference value, calculating the number of bubbles, and if the number of bubbles is not larger than a preset bubble threshold value, acquiring the cytopathology mounting.
Example nine
The present embodiment provides a further solution, wherein the step of using the digital scan image as an input of the digital pathology image evaluation model includes: acquiring qualified digital pathological images, and establishing a training data set of a proof sample; acquiring unqualified digital pathological images, and establishing a loading piece training data set; inputting a positive sample training data set and a negative sample training data set into a convolutional neural network model for training; and acquiring the trained convolutional neural network model as a digital pathological image evaluation model.
Example ten
The present embodiment provides a cytopathology image evaluating apparatus, as shown in fig. 12, including: the acquisition module is used for acquiring a cytopathology mounting, scanning the cytopathology mounting and acquiring a digital scanning image; an evaluation model module, configured to take the digital scan image as an input of the digital pathology image evaluation model, and output an evaluation result of the digital pathology image scan quality, where the evaluation result includes: a positive sample result, a result to be reproduced, and a result to be rescanned.
EXAMPLE eleven
The present embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of: obtaining a cytopathology mounting patch, scanning the cytopathology mounting patch and obtaining a digital scanning image; taking the digital scanning image as an input of a digital pathological image evaluation model, and outputting an evaluation result of the scanning quality of the digital pathological image, wherein the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
EXAMPLE twelve
The embodiment provides a computer device, which is shown in fig. 1 and includes a memory and a processor, wherein the memory stores a computer program, and the computer program, when executed by the processor, causes the processor to execute the following steps: obtaining a cytopathology mounting, scanning the cytopathology mounting and obtaining a digital scanning image; taking the digital scanning image as an input of a digital pathological image evaluation model, and outputting an evaluation result of the scanning quality of the digital pathological image, wherein the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
It should be noted that the embodiments of the cytopathology image evaluation method, the cytopathology image evaluation device, the storage medium, and the computer device described above are applicable to each other.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a non-volatile computer-readable storage medium, and can include the processes of the embodiments of the methods described above when the program is executed. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent application shall be subject to the appended claims.
Claims (10)
1. The cytopathology image evaluation method is characterized by comprising the following steps:
obtaining a cytopathology mounting, scanning the cytopathology mounting and obtaining a digital scanning image;
taking the digital scanning image as an input of a digital pathological image evaluation model, and outputting an evaluation result of the scanning quality of the digital pathological image, wherein the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
2. The cytopathology image evaluation method according to claim 1, wherein the step of obtaining a cytopathology envelope comprises:
acquiring the visual field of the cell pathological section by adjusting the coordinate parameters of the equipment;
analyzing the amount of cells in the field of view;
if the cell amount is larger than a preset cell amount threshold value, analyzing the staining quality of the cells in the visual field;
if the dyeing quality is larger than a preset RGB value, analyzing the number of bubbles in the visual field;
and if the number of the bubbles is not more than the preset bubble threshold value, obtaining the cytopathology mounting sheet.
3. The cytopathological image evaluation method of claim 2, wherein said step of scanning said cytopathological mount and acquiring a digital scan image comprises:
scanning the cytopathology mounting sheet to obtain a first image, analyzing the scanning definition of the first image, analyzing the scanning color of the first image if the scanning definition is greater than a preset definition threshold, and taking the first image as a digital scanning image if the scanning color is greater than a preset color threshold.
4. The cytopathology image evaluation method according to claim 3, wherein if the evaluation result is a positive sample result, a positive sample result is outputted;
if the evaluation result is the result to be reproduced, acquiring a second cytopathology mounting, scanning the second cytopathology mounting and acquiring a second image, taking the second image as the input of the digital pathology image evaluation model, and outputting the evaluation result of the second image;
and if the evaluation result is the result to be rescanned, rescanning the cytopathology mounting and acquiring a third image, taking the third image as the input of the digital pathology image evaluation model, and outputting the evaluation result of the third image.
5. The cytopathology image evaluation method according to claim 3, wherein if the evaluation result is a positive sample result, the positive sample result is analyzed as a cell quality evaluation result and an image quality evaluation result, and the cell quality evaluation result and the image quality evaluation result are input to the multidimensional evaluation model;
if the evaluation result is the result to be reproduced, acquiring a second cytopathology mounting, scanning the second cytopathology mounting and acquiring a second image, taking the second image as the input of the digital pathology image evaluation model, outputting the second evaluation result of the second image, analyzing the second evaluation result as a cell quality evaluation correction result, and taking the cell quality evaluation correction result as the input of the multi-dimensional evaluation model;
and if the evaluation result is the result to be rescanned, rescanning the cytopathology mounting and acquiring a third image, taking the third image as the input of the digital pathology image evaluation model, outputting the third evaluation result of the third image, analyzing the third evaluation result as an image quality evaluation correction result, and taking the image evaluation correction result as the input of the multi-dimensional evaluation model.
6. The cytopathology image evaluation method according to claim 5, wherein,
the step of using the cell quality evaluation correction result as an input of a multidimensional evaluation model includes:
analyzing the cell quality evaluation correction result, and outputting a pathological evaluation result of the multidimensional evaluation model according to the cell quality evaluation correction result, wherein the pathological evaluation result comprises a positive result and a negative result;
the step of inputting the image evaluation correction result as the input of the multidimensional evaluation model comprises the following steps:
analyzing the image evaluation and correction result, and outputting the image evaluation result of the multi-dimensional evaluation model according to the image evaluation and correction result, wherein the image evaluation result comprises the result of the image to be retrieved.
7. The cytopathology image evaluation method of claim 6, wherein the step of scanning the cytopathology mounting sheet to obtain a first image, analyzing a scanning resolution of the first image, analyzing a scanning color of the first image if the scanning resolution is greater than a preset resolution threshold, and regarding the first image as a digital scanning image if the scanning color is greater than a preset color threshold comprises:
scanning the cytopathology mounting sheet to obtain a first image, segmenting the first image into a plurality of first sub-images, and calculating the sub-definition of each first sub-image;
if the sub-definition of the first sub-image is greater than a preset sub-definition threshold, judging that the first sub-image is a clear first sub-image, and if the number of the first sub-images in the first image, of which the sub-definition is greater than the preset sub-definition threshold, reaches a preset number, then the scanning definition of the first image is greater than the preset definition threshold, and then the scanning color of the first image is analyzed;
and comparing the RGB parameters of the first image with preset RGB parameters, outputting an RGB parameter comparison result, and if the RGB parameter comparison result exceeds an RGB preset standard deviation range, taking the first image as a digital scanning image if the scanning color is larger than a preset color threshold value.
8. The cytopathology image evaluation method according to claim 7, wherein the step of obtaining a cytopathology envelope if the number of bubbles is not greater than a predetermined bubble threshold value comprises:
calculating the sub-definition difference value of the adjacent first sub-images in the first image, if the sub-definition difference value is larger than a preset sub-definition difference value, screening out a target area with the sub-definition difference value larger than the preset sub-definition difference value, calculating the number of bubbles, and if the number of bubbles is not larger than a preset bubble threshold value, acquiring the cytopathology mounting.
9. The cytopathology image evaluation method according to claim 1, wherein the step of inputting the digital scan image as a digital pathology image evaluation model comprises:
acquiring qualified digital pathological images, and establishing a proof training data set;
acquiring unqualified digital pathological images, and establishing a loading piece training data set;
inputting a positive sample training data set and a negative sample training data set into a convolutional neural network model for training;
and acquiring the trained convolutional neural network model as a digital pathological image evaluation model.
10. A cytopathology image evaluation device, comprising:
the acquisition module is used for acquiring a cytopathology mounting, scanning the cytopathology mounting and acquiring a digital scanning image;
the evaluation model module is used for taking the digital scanning image as the input of the digital pathological image evaluation model and outputting the evaluation result of the scanning quality of the digital pathological image, and the evaluation result comprises the following steps: a positive sample result, a result to be reproduced, and a result to be rescanned.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211502167.9A CN115880233A (en) | 2022-11-28 | 2022-11-28 | Cytopathology image evaluation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211502167.9A CN115880233A (en) | 2022-11-28 | 2022-11-28 | Cytopathology image evaluation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115880233A true CN115880233A (en) | 2023-03-31 |
Family
ID=85764322
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211502167.9A Pending CN115880233A (en) | 2022-11-28 | 2022-11-28 | Cytopathology image evaluation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115880233A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117765533A (en) * | 2024-02-22 | 2024-03-26 | 天津医科大学第二医院 | image processing method and system for oral mucosa cancer prediction |
-
2022
- 2022-11-28 CN CN202211502167.9A patent/CN115880233A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117765533A (en) * | 2024-02-22 | 2024-03-26 | 天津医科大学第二医院 | image processing method and system for oral mucosa cancer prediction |
| CN117765533B (en) * | 2024-02-22 | 2024-04-26 | 天津医科大学第二医院 | Image processing method and system for oral mucosa cancer prediction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111524137B (en) | Cell identification counting method and device based on image identification and computer equipment | |
| CN110110799A (en) | Cell sorting method, device, computer equipment and storage medium | |
| WO2021189771A1 (en) | Slide digitization information quality testing method and apparatus, and device and medium | |
| CN111753692B (en) | Target object extraction method, product detection method, device, computer and medium | |
| Wright et al. | The effect of quality control on accuracy of digital pathology image analysis | |
| CN113962976B (en) | Quality evaluation method for pathological slide digital image | |
| CN115201092B (en) | Method and device for acquiring cell scanning image | |
| CN112102247B (en) | Machine learning-based pathological section quality evaluation method and related equipment | |
| CN115880233A (en) | Cytopathology image evaluation method and device | |
| CN111161212B (en) | Digital pathological section mitosis image statistical method, device, equipment and medium | |
| CN109615613B (en) | Pathological section staining quality evaluation method and device, computer equipment and storage medium | |
| CN112464802B (en) | Automatic identification method and device for slide sample information and computer equipment | |
| CN111105407B (en) | Pathological section dyeing quality evaluation method, device, equipment and storage medium | |
| CN118115496B (en) | Wafer defect detection method and device | |
| CN114694143A (en) | Cell image recognition method and device based on optical means | |
| CN116959712B (en) | Lung adenocarcinoma prognosis method, system, equipment and storage medium based on pathological image | |
| CN117523211A (en) | Gray image processing method for butanediol propellant | |
| CN112233060B (en) | Screening method, device, equipment and medium for digital pathological image abnormal samples | |
| CN115359031A (en) | Digital pathological image slice quality evaluation method | |
| CN111681201B (en) | Image processing method, device, computer equipment and storage medium | |
| CN111048183A (en) | Method, device and equipment for homogenizing digital pathological image and storage medium | |
| CN117218139B (en) | Method and device for determining cell density of sample | |
| CN113850831B (en) | Plant ventilation tissue area measurement method and device and electronic equipment | |
| CN116152806B (en) | Bone marrow cell identification method and system based on convolutional neural network | |
| CN118097093B (en) | System for searching images on digital pathological section data set based on pathological large model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230703 Address after: 510005 No. 10, helix 3 Road, International Biological Island, Huangpu District, Guangzhou, Guangdong Applicant after: GUANGZHOU KINGMED CENTER FOR CLINICAL LABORATORY Applicant after: GUANGZHOU KINGMED DIAGNOSTICS GROUP Co.,Ltd. Address before: 510700 No. 10, helix 3 Road, International Biological Island, Huangpu District, Guangzhou City, Guangdong Province Applicant before: GUANGZHOU KINGMED CENTER FOR CLINICAL LABORATORY |
|
| TA01 | Transfer of patent application right |