CN118154975B - Tumor pathological diagnosis image classification method based on big data - Google Patents
Tumor pathological diagnosis image classification method based on big data Download PDFInfo
- Publication number
- CN118154975B CN118154975B CN202410364945.5A CN202410364945A CN118154975B CN 118154975 B CN118154975 B CN 118154975B CN 202410364945 A CN202410364945 A CN 202410364945A CN 118154975 B CN118154975 B CN 118154975B
- Authority
- CN
- China
- Prior art keywords
- pathological
- image
- lesion
- classification
- pathological change
- 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
- 238000010827 pathological analysis Methods 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 28
- 206010028980 Neoplasm Diseases 0.000 title claims abstract description 21
- 230000003902 lesion Effects 0.000 claims abstract description 99
- 230000036285 pathological change Effects 0.000 claims abstract description 95
- 231100000915 pathological change Toxicity 0.000 claims abstract description 95
- 230000001575 pathological effect Effects 0.000 claims abstract description 80
- 210000002307 prostate Anatomy 0.000 claims abstract description 33
- 230000007170 pathology Effects 0.000 claims abstract description 29
- 206010060862 Prostate cancer Diseases 0.000 claims abstract description 22
- 208000000236 Prostatic Neoplasms Diseases 0.000 claims abstract description 22
- 238000003745 diagnosis Methods 0.000 claims abstract description 14
- 210000001519 tissue Anatomy 0.000 claims description 8
- 230000002093 peripheral effect Effects 0.000 claims description 6
- 230000007704 transition Effects 0.000 claims description 5
- 108010075226 adenovirus factor PL Proteins 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 abstract description 3
- 238000004393 prognosis Methods 0.000 abstract description 3
- 230000011218 segmentation Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000002583 angiography Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 230000000762 glandular Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000002595 magnetic resonance imaging Methods 0.000 description 1
- 230000036210 malignancy Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008807 pathological lesion Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000002603 single-photon emission computed tomography Methods 0.000 description 1
- 238000011425 standardization method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000003325 tomography Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
- G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a tumor pathological diagnosis image classification method based on big data, relates to the technical field of image classification, and solves the technical problems that the classification is only carried out according to the lesion type and the lesion degree, but the positioning and classification of the prostatic cancer lesions are not considered to be very important for treatment selection and prognosis evaluation, so that doctors cannot quickly acquire similar tumor pathological diagnosis images according to the difference of lesion positions for reference; the method comprises the steps of obtaining a first classification, a second classification and a third classification corresponding to a pathological image through the pathological change type, the main pathological change position and the pathological change index of the pathological image, and classifying the pathological diagnosis image of the prostate cancer according to the positions of pathological change regions of the pathological diagnosis image in all parts of the prostate: the diagnosis images of the prostate cancer pathology are classified by combining the characteristics of pathology type, lesion position, lesion area size and the like, so that a more accurate diagnosis basis is provided for doctors.
Description
Technical Field
The invention relates to the technical field of image classification, in particular to a tumor pathological diagnosis image classification method based on big data.
Background
In recent years, with the continuous development of computer science and information technology, medical imaging technology has also been rapidly developed, and various new imaging devices are continuously emerging, such as Computed Tomography (CT), digital Subtraction Angiography (DSA), single photon emission tomography (SPECT), magnetic resonance imaging (MR I), and the like; based on the data of 1999-2006, prostate cancer is the most common malignancy in men with 15.3% of lifetime incidence, about 80% of prostate cancer patients present early diseases clinically limited to the prostate, in prostate diagnosis and treatment, along with the continuous development of medical technology, tumor pathological diagnosis images are increasingly widely applied in the medical field, doctors usually need to acquire similar prostate cancer pathological diagnosis images according to the patient images for reference diagnosis, and doctors are convenient for patients to carry out reference diagnosis;
The patent publication No. CN115953616A discloses a tissue pathology image classification device, which is applied to the technical field of image processing, the application provides a tissue pathology image classification device, an acquisition module acquires a pathology image to be tested firstly, then a call module calls a pre-trained pathology image classification network, an input module inputs the pathology image to be tested into the pathology image classification network for classification, and the construction and training steps of the pathology image classification network comprise: the method comprises the steps of obtaining a data set sample of a pathology image, wherein the data set sample comprises a pathology image set and a classification label set corresponding to the pathology image set, optimizing the pathology image set in the data set sample by adopting a data enhancement method and a data standardization method, and training a pathology image classification network through the optimized data set sample after an end-to-end pathology image classification network is constructed.
However, when classifying the pathological diagnosis image of the prostate cancer, the classification is generally performed only according to the lesion type and the lesion degree, however, the positioning and classification of the pathological lesions of the prostate cancer are not considered to be very important for treatment selection and prognosis evaluation, so that a doctor cannot quickly acquire similar pathological diagnosis images of the tumor according to the difference of the lesion positions for reference, and a method for classifying the pathological diagnosis image of the tumor based on big data is proposed based on the classification.
Disclosure of Invention
The invention aims to provide a tumor pathological diagnosis image classification method based on big data, which solves the technical problem that a doctor cannot quickly acquire similar tumor pathological diagnosis images for reference according to different lesion positions because the classification is only carried out according to the lesion types and the lesion degrees, but the positioning and classification of the prostatic cancer lesions are not considered to be very important for treatment selection and prognosis evaluation.
The aim of the invention can be achieved by the following technical scheme:
the tumor pathological diagnosis image classification method based on big data comprises the following steps of;
step one: acquiring a large amount of prostate cancer pathological diagnosis image data;
Step two: preliminary classification is carried out on the pathological diagnosis images according to different lesion types, and a first classification of the pathological diagnosis images is obtained;
step three: extracting a lesion region in a pathological diagnosis image;
step four: dividing a lesion area from the image;
Step five: marking a main lesion part and a secondary lesion part of the pathological image according to the pixel points corresponding to the lesion area, and taking the main lesion part and the secondary lesion part as a second classification;
step six: calculating a lesion index of the pathological image according to the lesion values of the main lesion part and the auxiliary lesion part;
step seven: and obtaining a third classification of the pathological image according to the pathological change index.
As a further scheme of the invention: the specific way to obtain the second classification is:
The method comprises the steps of obtaining pixel points corresponding to a pathological image, marking the pixel points as image pathological change values corresponding to the pathological image, obtaining pixel points of the pathological region in each part of the prostate, marking the pixel points as part pathological change values, marking the part with the part pathological change value larger than a preset value Y1 as a main pathological change part of the corresponding pathological image, and marking other parts as auxiliary pathological change parts.
As a further scheme of the invention: the pathological change index of the corresponding pathological image is obtained by the following specific modes:
Obtaining the product of the sum of the lesion values of the main lesion part and the preset fixed coefficient beta 1, obtaining the product of the sum of the lesion values of the auxiliary lesion part and the preset fixed coefficient beta 2, and finally marking the sum of the products as a lesion index.
As a further scheme of the invention: and obtaining pixel points of a pathological image pathological change region, marking the pixel points as image pathological change values, marking the pathological change value of the part of a main pathological change part as Wa, and marking the pathological change value of the part of a secondary pathological change part as Wb, wherein a+b=j, a is more than or equal to 1 and b is more than or equal to 1.
As a further scheme of the invention: the specific mode for extracting the pixel points of the pathological image pathological change area is as follows:
And (3) taking the pixel value of the lesion area as a foreground, assuming 1, and taking the pixel value of the normal tissue as a background, assuming 0, determining whether the pixel belongs to the lesion area by traversing each pixel of the image and checking whether the value is the foreground value, and extracting the pixel points of the lesion area.
As a further scheme of the invention: the third classification of the pathology image is obtained in the following specific manner:
When the pathological change index of the pathological change image is larger than a preset value Y2, the pathological change image is marked as an A-type pathological change image, when the pathological change index of the pathological change image is smaller than or equal to Y2 and larger than a preset value Y3, the pathological change image is marked as a B-type pathological change image, and when the pathological change index EB of the pathological change image is smaller than or equal to Y3, the pathological change image is marked as a C-type pathological change image.
As a further scheme of the invention: the collected large amount of prostate cancer pathological diagnosis image data is preprocessed before classifying the images, wherein the preprocessing operation comprises the operations of scaling, clipping and enhancing the images.
As a further scheme of the invention: each part of the prostate comprises a peripheral band, a central band and a transition band part, and the pathological image can show the peripheral band, the central band and the transition band parts of the prostate part corresponding to the diagnosis object;
As a further scheme of the invention: the method comprises the steps of obtaining lesion values of parts corresponding to pixel points of a lesion region of a pathological image in each part of a prostate, obtaining a numerical value number e which is not 0, marking the ratio between e and the total number of parts corresponding to the prostate as a part duty ratio R1, comprehensively analyzing the part duty ratio R1 of the pathological image and a lesion index EB, further obtaining a space occupation coefficient PL corresponding to the pathological image, and sequencing each pathological diagnosis image in the same class according to the space occupation coefficient PL corresponding to the pathological image from large to small.
As a further scheme of the invention: the specific mode for acquiring the occupation coefficient is as follows:
and obtaining the part ratio R1 of the pathological image through e/j=R1, and calculating to obtain the space factor PL corresponding to the pathological image through R1×β3+EB×β4=PL, wherein j refers to the number of parts corresponding to the prostate, j is greater than or equal to 1, and β3 and β4 are preset fixed factors.
The invention has the beneficial effects that:
(1) According to the invention, the first classification, the second classification and the third classification corresponding to the pathological image are obtained by classifying the pathological image of the pathological type, the main pathological part and the pathological index, and the pathological diagnosis image of the prostate cancer is classified according to the positions of pathological areas of the pathological diagnosis image in each part of the prostate: the pathological diagnosis images of the prostate cancer are classified by combining the characteristics of pathological types, pathological positions, pathological region sizes and the like, so that a more accurate diagnosis basis is provided for doctors, and the accuracy and the reliability of pathological diagnosis are improved;
(2) According to the invention, the ratio of the number of the non-0 numerical values in the pathological change values of the parts corresponding to the pixel points of each part of the pathological change region of the prostate to the total number of the parts corresponding to the prostate is used for obtaining the part ratio of the pathological change images, and each pathological diagnosis image in the same category is ordered according to the space occupation coefficient corresponding to the pathological change image, so that priority reference and calling are carried out on the pathological diagnosis images with more pathological change regions occupying the prostate part, more accurate diagnosis basis is further provided for doctors, and the accuracy and reliability of pathological diagnosis are improved.
Drawings
The invention is further described below with reference to the accompanying drawings.
FIG. 1 is a schematic diagram of a method framework of the present invention;
fig. 2 is a schematic diagram of a method for determining a third classification of a pathological image according to 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.
Example 1
Referring to fig. 1-2, the invention discloses a tumor pathological diagnosis image classification method based on big data, which comprises the following steps of;
Step one: acquiring a large amount of prostate cancer pathological diagnosis image data, wherein the prostate cancer pathological diagnosis image data comprises images of different lesion positions and different lesion areas, and the data can be acquired from public medical databases, hospitals, research institutions and the like;
Step two: preliminary classification is carried out on the pathological diagnosis images according to different pathological change types, so that pathological diagnosis image categories are obtained, and the pathological diagnosis image categories are used as first classification of pathological images;
The lesion types comprise glandular lesions, squamous cell lesions, small cell lesions and the like, and the existing SVM, decision tree classification algorithm or deep learning model can be used for comprehensively classifying the diagnosis images of the prostate cancer pathology;
Before classifying the images, a great amount of collected prostate cancer pathological diagnosis image data needs to be preprocessed, which includes operations such as image scaling, clipping, enhancing and the like, so that features in the images can be better extracted, and in addition, normalization processing is needed to be carried out on the images so as to map images with different sizes into the same space range, so that the method belongs to the existing and mature technology and is not repeated herein;
step three: extracting lesion areas of each pathological image in various pathological diagnosis images by using a target detection algorithm;
For each pathological image, a target detection algorithm can be used for detecting the position and the boundary of the pathological change of the image, such as a target detection method based on deep learning or a traditional edge detection algorithm, which belong to the existing and mature technology, so that the description is omitted here;
Step four: for the detected pathological region, the detected pathological region is segmented from the corresponding pathological image, so that the separation of the pathological region and surrounding normal tissues is ensured;
the specific segmentation mode is that an image segmentation algorithm, such as region-based segmentation or pixel-based segmentation, is adopted to separate a lesion region from surrounding normal tissues,
In the method based on the segmentation of the pixels, each pixel point is taken as a basic unit of the segmentation, the segmentation is realized by classifying or clustering the pixels, and then the segmentation is carried out on the prostate cancer pathological image by combining the image characteristics and the context information so as to ensure that the pathological change region and the surrounding normal tissues are accurately separated, thus the method belongs to the existing and mature technology and is not repeated herein;
Step five: extracting the segmented lesion areas, marking the main lesion areas of the pathological images in each part of the prostate according to the pixel points corresponding to the lesion areas, and taking the main lesion areas as the second classification of the pathological images, wherein the specific mode is as follows:
various parts of the prostate including peripheral, central and transitional zones, which can be determined by pathological section or MR I images; the pathological image should be able to show the peripheral, central and transition zones of the prostate part of the subject to be diagnosed;
obtaining pixel points corresponding to a pathological region, marking the pixel points as image pathological change values corresponding to pathological images, obtaining pixel points of the pathological region in each part of the prostate, marking the pixel points as part pathological change values, marking the parts with the part pathological change values larger than a preset value as main pathological change parts of the corresponding pathological images, and marking the other parts as auxiliary pathological change parts;
obtaining pixel points of pathological image pathological change areas, marking the pixel points as image pathological change values, expressing the image pathological change values through BA, obtaining pixel points of pathological change areas in all parts of the prostate, marking the pixel points as part pathological change values Wj, wherein j refers to the number of parts corresponding to the prostate, j is more than or equal to 1, marking the parts with the part pathological change values Wj larger than a preset value Y1 as main pathological change parts of the corresponding pathological images, and marking the rest parts as auxiliary pathological change parts;
The size of the lesion area is represented by the pixel points of the lesion area, so that the later judgment and analysis of the lesion area are facilitated, and meanwhile, the size of the lesion area in each part of the prostate is represented by the pixel points, so that the subsequent analysis of the lesion area in each part of the prostate is facilitated;
step six: according to the lesion values of the parts corresponding to the main lesion part and the auxiliary lesion part respectively, the lesion index of the corresponding pathological image is obtained, and the specific mode is as follows:
marking the lesion value of the main lesion part as Wa, and marking the lesion value of the auxiliary lesion part as Wb, wherein a+b=j, a is more than or equal to 1 and b is more than or equal to 1, and the specific value of the preset value Y1 is drawn up by related staff according to experience;
obtaining the product between the sum of the lesion values of the main lesion part and a preset fixed coefficient beta 1, obtaining the product between the sum of the lesion values of the auxiliary lesion part and a preset fixed coefficient beta 2, and finally marking the sum of the products as a lesion index;
I.e. by the formula Calculating to obtain a pathology index EB of a corresponding pathology image, wherein beta 1 and beta 2 are preset fixed coefficients, specific numerical values are drawn up by related staff according to experience, a is more than or equal to e and more than or equal to 1, and b is more than or equal to g is more than or equal to 1;
The specific mode for extracting the pixel points of the pathological image pathological change area is as follows: determining whether the pixel value of the lesion area is a foreground, assuming that the pixel value of the lesion area is 1, and assuming that the pixel value of the normal tissue is 0, and extracting the pixel points of the lesion area by traversing each pixel of the image and checking whether the value of each pixel is a foreground value;
step seven, according to the pathological change index of the pathological image, obtaining a third classification of the pathological image, wherein the specific mode is as follows:
When the pathological change index EB of the pathological change image is larger than a preset value Y2, marking as an A-type pathological change image, when the pathological change index EB of the pathological change image is smaller than or equal to Y2 and larger than a preset value Y3, marking as a B-type pathological change image, and when the pathological change index EB of the pathological change image is smaller than or equal to Y3, marking as a C-type pathological change image, so as to obtain a third classification of the pathological change image;
The method comprises the steps of performing preliminary classification on a pathological diagnosis image according to different pathological change types, obtaining a first classification of the pathological diagnosis image, marking a main pathological change position and a secondary pathological change position of the pathological diagnosis image according to pixel points corresponding to pathological change regions, completing second classification of the pathological diagnosis image, obtaining a third classification of the pathological change type, the main pathological change position and the pathological change index of the pathological diagnosis image according to pathological change indexes, obtaining the first classification, the second classification and the third classification corresponding to the pathological change image, and classifying the pathological diagnosis image of the prostate cancer according to positions of the pathological change regions of the pathological diagnosis image in all positions of the prostate: the diagnosis images of the prostate cancer pathology are classified by combining the characteristics of pathology type, pathology position, pathology area size and the like, so that a more accurate diagnosis basis is provided for doctors, and the accuracy and reliability of the pathology diagnosis are improved.
Example two
As an embodiment two of the present application, when the present application is implemented, compared with the embodiment one, the technical solution of the present embodiment is different from the embodiment one only in that in the present embodiment,
Obtaining part lesion values Wj corresponding to pixel points of a pathological image lesion area in each part of the prostate, obtaining numerical value number e which is not 0, marking the ratio between e and the total number j of parts corresponding to the prostate as part duty ratio R1, comprehensively analyzing the part duty ratio R1 of the pathological image and a lesion index EB, and further obtaining a duty factor PL corresponding to the pathological image;
the part ratio R1 of the pathological image is obtained through e/j=R1, and then the corresponding space occupation coefficient PL of the pathological image is obtained through R1 xβ3+EBxβ4=PL in a calculation mode, wherein β3 and β4 are preset fixed coefficients, and specific numerical values are drawn up by related staff according to experience;
Ordering each pathological diagnosis image in the same category according to the corresponding occupation coefficient PL of the pathological image, and ordering from big to small according to the corresponding value of the occupation coefficient PL;
The ratio of the number of the non-0 numerical values in the pathological change values of the parts corresponding to the pixel points of each part of the pathological change region of the prostate to the total number of the parts corresponding to the prostate is used for obtaining the part duty ratio of the pathological images, and each pathological diagnosis image in the same category is ordered according to the duty ratio corresponding to the pathological image, so that priority reference and calling are carried out on the pathological diagnosis images with pathological change regions occupying more parts of the prostate, more accurate diagnosis basis is further provided for doctors, and the accuracy and reliability of pathological diagnosis are improved.
Example III
As an embodiment three of the present application, in the implementation of the present application, the technical solution of the present embodiment is to combine the solutions of the above embodiment one and embodiment two compared with the embodiment one and embodiment two.
The above formulas are all formulas with dimensionality removed and numerical calculation, the formulas are formulas with the latest real situation obtained by software simulation through collecting a large amount of data, and preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.
The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (8)
1. The tumor pathological diagnosis image classification method based on big data is characterized by comprising the following steps of;
step one: acquiring a large amount of prostate cancer pathological diagnosis image data;
Step two: preliminary classification is carried out on the pathological diagnosis images according to different lesion types, and a first classification of the pathological diagnosis images is obtained;
step three: extracting a lesion region in a pathological diagnosis image;
step four: dividing a lesion area from the image;
Step five: marking a main lesion part and a secondary lesion part of the pathological image according to the pixel points corresponding to the lesion area, and taking the main lesion part and the secondary lesion part as a second classification;
step six: calculating a lesion index of the pathological image according to the lesion values of the main lesion part and the auxiliary lesion part;
Step seven: acquiring a third classification of the pathological image according to the pathological change index;
The pathological change index of the corresponding pathological image is obtained by the following specific modes:
obtaining the product between the sum of the lesion values of the main lesion part and a preset fixed coefficient beta 1, obtaining the product between the sum of the lesion values of the auxiliary lesion part and a preset fixed coefficient beta 2, and finally marking the sum of the products as a lesion index;
the third classification of the pathology image is obtained in the following specific manner:
When the pathological change index of the pathological image is larger than a preset value Y2, the pathological change image is marked as an A-type pathological change image, when the pathological change index of the pathological image is smaller than or equal to Y2 and larger than a preset value Y3, the pathological change image is marked as a B-type pathological change image, and when the pathological change index EB of the pathological image is smaller than or equal to the preset value Y3, the pathological change image is marked as a C-type pathological change image, so that a third classification of the pathological image is obtained.
2. The method for classifying tumor pathological diagnosis images based on big data according to claim 1, wherein the specific way of obtaining the second classification is:
The method comprises the steps of obtaining pixel points corresponding to a pathological image, marking the pixel points as image pathological change values corresponding to the pathological image, obtaining pixel points of the pathological region in each part of the prostate, marking the pixel points as part pathological change values, marking the part with the part pathological change value larger than a preset value Y1 as a main pathological change part of the corresponding pathological image, and marking other parts as auxiliary pathological change parts.
3. The method for classifying tumor pathological diagnosis images based on big data according to claim 1, wherein pixels of pathological image lesion areas are obtained, marked as image lesion values corresponding to pathological images, the site lesion values of main lesion sites are marked as Wa, the site lesion values of secondary lesion sites are marked as Wb, wherein a+b=j, and a is equal to or greater than 1 and b is equal to or greater than 1.
4. The method for classifying tumor pathological diagnosis images based on big data according to claim 2, wherein the specific way of extracting the pixels of the pathological image lesion area is as follows:
and (3) taking the pixel value of the lesion area as a foreground, assuming 1, and taking the pixel value of the normal tissue as a background, assuming 0, determining whether the pixel belongs to the lesion area by traversing each pixel of the image and checking whether the value is the foreground value, and extracting the pixel points of the lesion area.
5. The method of classifying tumor pathology diagnostic image based on big data according to claim 1, wherein the collected large amount of prostate cancer pathology diagnostic image data is preprocessed before classifying the image, the preprocessing operation including scaling, cropping, and enhancing operations of the image.
6. The method of classifying tumor pathology diagnostic images based on big data according to claim 2, wherein each part of the prostate comprises peripheral, central and transition zone parts, and the pathology image should be able to show the peripheral, central and transition zone parts of the prostate part corresponding to the diagnosis subject.
7. The method for classifying tumor pathological diagnosis images based on big data according to claim 1, wherein the pathological change values of the pathological change areas of the pathological images corresponding to the pixel points in each part of the prostate are obtained, the number e of the pixels which is not 0 is obtained, the ratio between e and the total number of the parts corresponding to the prostate is marked as a part duty ratio R1, the part duty ratio R1 of the pathological images and a pathological change index EB are comprehensively analyzed, the space occupation coefficient PL corresponding to the pathological images is obtained, the pathological diagnosis images in the same classification are ranked according to the space occupation coefficient PL corresponding to the pathological images from big to small.
8. The method for classifying tumor pathological diagnosis images based on big data according to claim 7, wherein the specific way of obtaining the occupancy coefficients is:
and obtaining the part ratio R1 of the pathological image through e/j=R1, and calculating to obtain the space factor PL corresponding to the pathological image through R1×β3+EB×β4=PL, wherein j refers to the number of parts corresponding to the prostate, j is greater than or equal to 1, and β3 and β4 are preset fixed factors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410364945.5A CN118154975B (en) | 2024-03-27 | 2024-03-27 | Tumor pathological diagnosis image classification method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410364945.5A CN118154975B (en) | 2024-03-27 | 2024-03-27 | Tumor pathological diagnosis image classification method based on big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118154975A CN118154975A (en) | 2024-06-07 |
| CN118154975B true CN118154975B (en) | 2024-10-01 |
Family
ID=91293405
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410364945.5A Active CN118154975B (en) | 2024-03-27 | 2024-03-27 | Tumor pathological diagnosis image classification method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118154975B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118762246B (en) * | 2024-09-09 | 2024-11-22 | 大连壹致科技有限公司 | An intelligent classification method for ENT lesion images |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082490A (en) * | 2022-08-23 | 2022-09-20 | 腾讯科技（深圳）有限公司 | Anomaly prediction method, and training method, device and equipment of anomaly prediction model |
| CN116524315A (en) * | 2022-01-17 | 2023-08-01 | 山东大学 | A Method for Recognition and Segmentation of Lung Cancer Pathological Tissue Slices Based on Mask R-CNN |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105975793A (en) * | 2016-05-23 | 2016-09-28 | 麦克奥迪（厦门）医疗诊断系统有限公司 | Auxiliary cancer diagnosis method based on digital pathological images |
| CN111968091B (en) * | 2020-08-19 | 2022-04-01 | 南京图格医疗科技有限公司 | Method for detecting and classifying lesion areas in clinical image |
| CN113889261B (en) * | 2021-09-23 | 2022-06-10 | 之江实验室 | Pathological feature assistance-based PET/CT automatic lung cancer diagnosis classification model training method |
-
2024
- 2024-03-27 CN CN202410364945.5A patent/CN118154975B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116524315A (en) * | 2022-01-17 | 2023-08-01 | 山东大学 | A Method for Recognition and Segmentation of Lung Cancer Pathological Tissue Slices Based on Mask R-CNN |
| CN115082490A (en) * | 2022-08-23 | 2022-09-20 | 腾讯科技（深圳）有限公司 | Anomaly prediction method, and training method, device and equipment of anomaly prediction model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118154975A (en) | 2024-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109583440B (en) | Medical image auxiliary diagnosis method and system combining image recognition and report editing | |
| Mahalakshmi et al. | Detection of brain tumor by particle swarm optimization using image segmentation | |
| Bai et al. | Automatic segmentation of cervical region in colposcopic images using K-means | |
| CN114782307A (en) | Deep learning-based enhanced CT image rectal cancer staging auxiliary diagnosis system | |
| CN114974575A (en) | Breast cancer neoadjuvant chemotherapy curative effect prediction device based on multi-feature fusion | |
| CN110728239B (en) | An automatic recognition system for gastric cancer enhanced CT images using deep learning | |
| Xu et al. | Using transfer learning on whole slide images to predict tumor mutational burden in bladder cancer patients | |
| CN118154975B (en) | Tumor pathological diagnosis image classification method based on big data | |
| CN114529505A (en) | Breast lesion risk assessment system based on deep learning | |
| Florez et al. | Emergence of radiomics: novel methodology identifying imaging biomarkers of disease in diagnosis, response, and progression | |
| Song et al. | Prostate lesion segmentation based on a 3D end-to-end convolution neural network with deep multi-scale attention | |
| CN116580007A (en) | Full-breast information analysis method for predicting curative effect of new adjuvant therapy of breast cancer | |
| CN112071418B (en) | Gastric cancer peritoneal metastasis prediction system and method based on enhanced CT image histology | |
| CN118657756A (en) | Intelligent decision-making support system and method for nursing care of patients with brain tumors | |
| Jaffar et al. | GA and morphology based automated segmentation of lungs from Ct scan images | |
| Sreelatha et al. | A survey work on early detection methods of melanoma skin cancer | |
| Lai et al. | Interstitial fibrosis and tubular atrophy measurement via hierarchical extractions of kidney and atrophy regions with deep learning method | |
| KR101162599B1 (en) | An automatic detection method of Cardiac Cardiomegaly through chest radiograph analyses and the recording medium thereof | |
| Li et al. | A dual attention-guided 3d convolution network for automatic segmentation of prostate and tumor | |
| Junior et al. | Evaluating margin sharpness analysis on similar pulmonary nodule retrieval | |
| Touil et al. | A new conditional region growing approach for microcalcification delineation in mammograms | |
| CN116705295A (en) | Diagnosis and evaluation method after endometrial cancer conservation treatment based on pathology deep learning | |
| Priyadharshini et al. | Artificial Intelligence Assisted Improved Design to Predict Brain Tumor on Earlier Stages using Deep Learning Principle | |
| Gemme et al. | Automatic MPST-cut for segmentation of carpal bones from MR volumes | |
| Zang et al. | Tuberculin skin test result detection method based on CSN-II and improved OTSU method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |