CN112435243A - Automatic analysis system and method for full-slice digital pathological image - Google Patents
Automatic analysis system and method for full-slice digital pathological image Download PDFInfo
- Publication number
- CN112435243A CN112435243A CN202011346662.6A CN202011346662A CN112435243A CN 112435243 A CN112435243 A CN 112435243A CN 202011346662 A CN202011346662 A CN 202011346662A CN 112435243 A CN112435243 A CN 112435243A
- Authority
- CN
- China
- Prior art keywords
- image
- full
- pathological
- small blocks
- pathological image
- 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
- 230000001575 pathological effect Effects 0.000 title claims abstract description 142
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000004458 analytical method Methods 0.000 title claims abstract description 52
- 238000013136 deep learning model Methods 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 35
- 230000007170 pathology Effects 0.000 claims abstract description 33
- 230000011218 segmentation Effects 0.000 claims abstract description 33
- 238000001514 detection method Methods 0.000 claims abstract description 32
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims description 25
- 230000008569 process Effects 0.000 claims description 13
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 230000002708 enhancing effect Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 238000003709 image segmentation Methods 0.000 claims description 5
- 238000002372 labelling Methods 0.000 claims description 5
- 230000003902 lesion Effects 0.000 claims description 3
- 238000012800 visualization Methods 0.000 claims description 3
- 238000003745 diagnosis Methods 0.000 abstract description 14
- 238000010606 normalization Methods 0.000 abstract description 9
- 230000000007 visual effect Effects 0.000 abstract 1
- 210000001519 tissue Anatomy 0.000 description 33
- 238000013135 deep learning Methods 0.000 description 19
- 206010028980 Neoplasm Diseases 0.000 description 15
- 238000012360 testing method Methods 0.000 description 10
- 201000011510 cancer Diseases 0.000 description 9
- 238000013145 classification model Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- WZUVPPKBWHMQCE-UHFFFAOYSA-N Haematoxylin Chemical compound C12=CC(O)=C(O)C=C2CC2(O)C1C1=CC=C(O)C(O)=C1OC2 WZUVPPKBWHMQCE-UHFFFAOYSA-N 0.000 description 6
- 210000003855 cell nucleus Anatomy 0.000 description 6
- 230000012010 growth Effects 0.000 description 6
- 230000000903 blocking effect Effects 0.000 description 5
- 239000004530 micro-emulsion Substances 0.000 description 5
- 239000001993 wax Substances 0.000 description 5
- 208000009956 adenocarcinoma Diseases 0.000 description 4
- 230000004069 differentiation Effects 0.000 description 4
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 238000004043 dyeing Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000010191 image analysis Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 208000010507 Adenocarcinoma of Lung Diseases 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 239000003086 colorant Substances 0.000 description 3
- 238000004195 computer-aided diagnosis Methods 0.000 description 3
- YQGOJNYOYNNSMM-UHFFFAOYSA-N eosin Chemical compound [Na+].OC(=O)C1=CC=CC=C1C1=C2C=C(Br)C(=O)C(Br)=C2OC2=C(Br)C(O)=C(Br)C=C21 YQGOJNYOYNNSMM-UHFFFAOYSA-N 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 201000005249 lung adenocarcinoma Diseases 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 210000004940 nucleus Anatomy 0.000 description 3
- 238000013526 transfer learning Methods 0.000 description 3
- 206010006187 Breast cancer Diseases 0.000 description 2
- 208000026310 Breast neoplasm Diseases 0.000 description 2
- 230000001464 adherent effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 210000004027 cell Anatomy 0.000 description 2
- 230000018044 dehydration Effects 0.000 description 2
- 238000006297 dehydration reaction Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 201000010198 papillary carcinoma Diseases 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000005406 washing Methods 0.000 description 2
- 241000321096 Adenoides Species 0.000 description 1
- 201000009030 Carcinoma Diseases 0.000 description 1
- 101000583175 Homo sapiens Prolactin-inducible protein Proteins 0.000 description 1
- 208000037396 Intraductal Noninfiltrating Carcinoma Diseases 0.000 description 1
- 206010073099 Lobular breast carcinoma in situ Diseases 0.000 description 1
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 1
- 208000007433 Lymphatic Metastasis Diseases 0.000 description 1
- 208000007054 Medullary Carcinoma Diseases 0.000 description 1
- 206010027476 Metastases Diseases 0.000 description 1
- 102100030350 Prolactin-inducible protein Human genes 0.000 description 1
- 206010041067 Small cell lung cancer Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 210000002534 adenoid Anatomy 0.000 description 1
- 208000002517 adenoid cystic carcinoma Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 201000008275 breast carcinoma Diseases 0.000 description 1
- 201000005389 breast carcinoma in situ Diseases 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000002380 cytological effect Effects 0.000 description 1
- 210000000805 cytoplasm Anatomy 0.000 description 1
- 230000001086 cytosolic effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 208000028715 ductal breast carcinoma in situ Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000007773 growth pattern Effects 0.000 description 1
- 238000003703 image analysis method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000008595 infiltration Effects 0.000 description 1
- 238000001764 infiltration Methods 0.000 description 1
- 206010073095 invasive ductal breast carcinoma Diseases 0.000 description 1
- 201000010985 invasive ductal carcinoma Diseases 0.000 description 1
- 206010073096 invasive lobular breast carcinoma Diseases 0.000 description 1
- 208000003849 large cell carcinoma Diseases 0.000 description 1
- 201000011059 lobular neoplasia Diseases 0.000 description 1
- 201000005202 lung cancer Diseases 0.000 description 1
- 208000020816 lung neoplasm Diseases 0.000 description 1
- 201000005243 lung squamous cell carcinoma Diseases 0.000 description 1
- 210000004698 lymphocyte Anatomy 0.000 description 1
- 230000036210 malignancy Effects 0.000 description 1
- 208000023356 medullary thyroid gland carcinoma Diseases 0.000 description 1
- 230000009401 metastasis Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 201000010879 mucinous adenocarcinoma Diseases 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 208000002154 non-small cell lung carcinoma Diseases 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 239000012188 paraffin wax Substances 0.000 description 1
- 238000010827 pathological analysis Methods 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 201000007416 salivary gland adenoid cystic carcinoma Diseases 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 208000000587 small cell lung carcinoma Diseases 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 206010041823 squamous cell carcinoma Diseases 0.000 description 1
- 238000010186 staining Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000011269 treatment regimen Methods 0.000 description 1
- 201000007423 tubular adenocarcinoma Diseases 0.000 description 1
- 210000004881 tumor cell Anatomy 0.000 description 1
- 208000029729 tumor suppressor gene on chromosome 11 Diseases 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Abstract
The present disclosure provides an automatic analysis system and method for a full-slice digital pathology image, comprising acquiring a full-slice digital pathology image; carrying out block processing on the full-slice digital pathological image; dividing the pathological image block into a background blank area and a foreground tissue area, and performing color normalization on the image block in the foreground area to obtain small blocks which can be directly used for analyzing the foreground tissue area; after image preprocessing, classifying the pathological image blocks to obtain a prediction subtype classification result, a segmentation result and a target detection result of each pathological image block; after the full-slice digital pathological image is obtained, the full-slice digital pathological image is partitioned and preprocessed, so that image classification, segmentation and target detection tasks can be performed by using a deep learning model, and then visual display is performed, so that a pathologist is assisted in observing and analyzing the image, and the diagnosis efficiency and the diagnosis accuracy are improved.
Description
Technical Field
The disclosure relates to the technical field of medical image processing, in particular to an automatic analysis system and method for full-slice digital pathological images.
Background
Histopathological examination is a pathomorphological method for examining pathological changes in tissues to analyze and diagnose diseases, and is currently the most accurate method for diagnosing cancer. The pathological images can help doctors analyze the disease condition of patients to obtain the specific conditions of tumor cells, such as differentiation degree, lymph node metastasis and the like, and are helpful for disease diagnosis, staging and prognosis. Typically, a pathologist classifies tissues and receives histopathological reports to determine the next treatment regimen. The pathological specimen is shot by adopting the scanning image equipment, and a high-quality pathological digital image can be obtained.
In recent years, with the rapid development of artificial intelligence technology, Computer Aided Diagnosis (CAD) has been successful in the medical field. The main methods of CAD in the diagnosis of pathological images include conventional machine learning and the more popular deep learning in recent years. The analysis effect of the method mainly depends on the effect of the previous manual feature extraction. Compared with the traditional method, the deep learning does not need manual feature extraction, deep features of pathological images can be automatically mined, and end-to-end optimization is directly carried out; the deep learning method is used for realizing medical image classification, target area detection, target area segmentation and retrieval, and assisting a doctor to finish analysis and diagnosis work; the artificial intelligence auxiliary diagnosis has the advantages of high efficiency, objectivity and stability.
The inventor finds that the current pathological image analysis still has the following problems: the diagnosis of doctors needs to take a long time, and the requirement on the professional ability of the doctors is high; the traditional machine learning method analyzes pathological images and mainly depends on the effect of feature extraction, which has higher requirements on professional knowledge of research personnel; the existing method based on deep learning can only analyze small-size images, but the size of the digital pathological images is too large, and the existing method cannot be directly used for analyzing large-size full-slice digital pathological images.
Disclosure of Invention
In order to solve the above problems, the present disclosure provides an automatic analysis system and method for a full-slice digital pathological image, which can assist a pathologist in observing an analysis image, and improve diagnosis efficiency and diagnosis accuracy.
According to a first aspect of embodiments of the present disclosure, there is provided an automatic analysis system of a full-slice digital pathology image, comprising:
the image preprocessing module is used for dividing the acquired full-slice digital pathological image into a plurality of image small blocks; dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks; performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and the automatic analysis module is used for analyzing and processing the processed image small blocks by utilizing the pre-trained deep learning model and realizing the analysis of the full-slice pathological image according to the processing result.
Further, the analysis of the full-slice digital pathological image is realized according to the analysis processing result, and the specific steps are that the prediction result of the automatic analysis module is marked on each pathological image small block, and the marked image small blocks are spliced back to the size of the original input image.
Furthermore, the spliced image is displayed on a display, so that a user can assist in judging the position and the type of a lesion in the image according to a marking result.
According to a second aspect of the embodiments of the present disclosure, there is provided a method for automatically analyzing a full-slice digital pathology image, including:
dividing the acquired full-slice digital pathological image into a plurality of image small blocks;
dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks;
performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and analyzing the processed image small blocks by utilizing a pre-trained deep learning model, and analyzing the full-slice digital pathological image according to an analysis processing result.
According to a third aspect of the embodiments of the present disclosure, there is provided an electronic device, including a memory, a processor and a computer program stored in the memory for execution, wherein the processor implements the method for automatically analyzing a full-slice digital pathological image.
According to a fourth aspect of the embodiments of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method for automatic analysis of a full-slice digital pathology image.
Compared with the prior art, the beneficial effect of this disclosure is:
(1) in terms of completeness, the present disclosure provides a full-slice digital pathological image processing method and process based on a deep learning model for the first time, and provides a complete process for performing digital pathological image processing by using a deep learning method, including: acquiring a full-slice digital pathological image; carrying out block pretreatment on the full-slice pathological image, and dividing the full-slice pathological image into small digital pathological image blocks which can be used for direct analysis; after obtaining small blocks of the pathological image, classifying each image block to obtain subtype classification results; or after obtaining the pathological image small blocks, carrying out segmentation processing on each image block to obtain subtype segmentation results; or after obtaining the pathological image small blocks, detecting each image block to obtain a subtype detection result. In addition, the present disclosure provides a full-slice pathological image corresponding label processing flow; a procedure for visualizing results.
(2) In the practicability and the expansibility, the method provided by the disclosure can be used for classifying and segmenting the digital pathological image sub-regions and detecting the target by the trained deep learning model, so that the subtype of the full-slice pathological image is marked in advance, a pathologist can be assisted to observe and analyze the image, and the diagnosis efficiency and the diagnosis accuracy are improved. Meanwhile, the method provides a complete digital pathological image analysis process, can complete classification, segmentation and target detection of pathological image sub-regions, and can be fused with other pathological image analysis methods.
(3) On the aspects of calculation efficiency and operation speed, the method is based on the deep learning model, and the trained model can realize the operation analysis of the pathological image block through one-time forward transmission, so that the operation analysis of the full-slice pathological image can be realized in a short time.
Advantages of additional aspects of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.
Fig. 1 is a schematic flowchart of a full-slice digital pathology image processing method based on deep learning according to an embodiment of the present disclosure;
FIG. 2 is a flowchart of acquiring a full-slice digital pathology image according to a first embodiment of the disclosure;
FIG. 3 is a flow chart of pre-processing a full-slice digital pathology image according to an embodiment of the present disclosure;
FIG. 4(a) is a diagram of an example of a full-slice digital pathology image in accordance with an embodiment of the present disclosure;
FIG. 4(b) is a diagram of another example of a full-slice digital pathology image in accordance with one embodiment of the present disclosure;
fig. 5 is a diagram of an example of a full-slice digital pathology image after being diced in accordance with an embodiment of the present disclosure;
FIG. 6 is a flow chart of processing a Label corresponding to a full-slice digital pathology image according to an embodiment of the disclosure;
FIG. 7 is a flowchart of a test image and visualization process according to an embodiment of the disclosure;
fig. 8 is a diagram illustrating an example of classification results for all-slice digital pathology images according to an embodiment of the present disclosure.
The specific implementation mode is as follows:
the present disclosure is further described with reference to the following drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
It is noted that the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of methods and systems according to various embodiments of the present disclosure. It should be noted that each block in the flowchart or block diagrams may represent a module, a segment, or a portion of code, which may comprise one or more executable instructions for implementing the logical function specified in the respective embodiment. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The first embodiment is as follows:
the embodiment aims to provide an automatic analysis system for a full-slice digital pathological image.
An automated analysis system for full-slice digital pathology images, comprising:
an image preprocessing module: the system comprises a data acquisition unit, a data processing unit and a data processing unit, wherein the data acquisition unit is used for acquiring a full-slice digital pathological image; dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks; performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
the image blocks are named according to a certain naming rule and used for splicing and restoring the processed images.
An automatic analysis module: the method is used for analyzing and processing the processed image small blocks by utilizing the pre-trained deep learning model and analyzing the full-section pathological image according to the processing result.
Wherein the analysis of the processed image patches by the deep learning model comprises:
predicting the image small blocks by using the trained deep learning model to obtain a prediction subtype classification result of each pathological image small block;
specifically, the deep learning model classified by using sample data may be trained in advance to obtain a trained deep learning image classification model, where the sample data includes a digital pathological image and a subtype label corresponding to the pathological image. The training process of the network model can be realized by the existing deep learning training method without specific limitation. After the trained deep learning classification model is obtained, the processor uses the trained model to classify each pathological image block to obtain a classification result of each pathological image block, and the classification result is displayed on the pathological image block.
Or predicting the image small blocks by using the trained deep learning model to obtain a prediction subtype segmentation result of each pathological image small block;
specifically, the segmented deep learning model may be trained in advance using sample data, so as to obtain a trained deep learning image segmentation model, where the sample data includes a digital pathological image and a subtype label corresponding to the pathological image. The training process of the network model can be realized by the existing deep learning training method without specific limitation. After the trained deep learning segmentation model is obtained, the processor performs segmentation processing on each pathological image block by using the trained model to obtain a segmentation result corresponding to each pathological image block, and displays the segmentation result on the pathological image block.
Or predicting the image small blocks by using the trained deep learning model to obtain a predicted subtype target detection result of each pathological image small block;
specifically, the deep learning model for target detection may be trained in advance using sample data to obtain a trained deep learning image target detection model, where the sample data includes a digital pathological image and a target detection label corresponding to the pathological image. The training process of the network model can be realized by the existing deep learning training method without specific limitation. After the trained deep learning target detection model is obtained, the processor uses the trained model to perform detection processing on each pathological image block to obtain the detection result of each pathological image block, and the target detection result is displayed on the pathological image block.
A visualization module: the prediction result is displayed on each pathological image patch, and the marked patch is spliced back to the size of the original input image.
Further, the training process of the deep learning model comprises: a model training phase and a model testing phase.
In the model training stage, the deep learning model is trained by using the sample data to obtain the trained deep learning model. In the training stage, a system analysis task, model selection, a model use data path, training set and verification set quantity, a block side length, a task type, a model storage path, model training parameters such as initialization, deviation, regularization, a learning rate, an optimization algorithm, iteration times and the like need to be set, so that model training is realized. Optionally, the number of categories and the names of the corresponding subtypes in different tasks, and specific parameters to be set for different tasks.
And the model test stage comprises a model input module, a task type module and an output module.
The input comprises a path of a test folder, a test model path, a test model, the number of test images and a test result output path. Optionally, in the subtype classification task, the category and the corresponding display color, transparency, and the like. Optionally, in the subtype division task, categories and corresponding display colors, transparency, and the like. Optionally, in the subtype detection task, the confidence threshold and the border corresponding to the category display colors.
The system is mainly used for an automatic analysis system of digital pathological images, and comprises subtype classification, subtype segmentation and subtype detection. In use, only the model testing phase needs to be performed when using the trained model to analyze pathology images. In order to enable the system to be applied to different histopathology image analysis and enhance the expandability of the system, the system is added with a model training stage.
Specifically, as shown in fig. 1, the specific processing steps of the automatic analysis system of the present disclosure are as follows, and the application of the process to the full-slice digital histopathological image analysis is explained as an example, and the process includes the following steps:
step 1: a full-slice digital pathology image is acquired.
Specifically, as shown in fig. 2, a flow of acquiring a full-slice digital pathological image is obtained, and the cut pathological tissue is wrapped in a wax block; pathological tissue wax block is sliced, H & E (H is the abbreviation of hematoxylin, the stained cell nucleus is blue, E is the abbreviation of eosin, the stained cytoplasm is red) is stained, then the stained pathological tissue wax block is scanned into a computer by a digital slice scanner, and the scanned full-slice digital pathological image can be amplified and reduced, has higher resolution ratio, and can be used for extracting and analyzing the histological and cytological characteristics of pathological tissues by a computer. The embodiment adopts a conventional mode of acquiring a full-slice digital pathological image, and the specific mode is not modified.
Step 2: the full-slice digital pathological image is processed in a blocking mode and divided into small image blocks which can be used for direct analysis.
In particular, the blocking processing of the full-slice digital pathological image refers to the division of the image into a plurality of image small blocks which can be used for analysis. In this embodiment, an OpenSlide software package is used to read a digital pathological image file, and the size of a block, the size of an overlap, whether an edge is discarded, and the like are set, so that the digital pathological image block is realized.
And step 3: and (5) image preprocessing.
The pathological image preprocessing step comprises two steps of dividing the pathological image block into a background blank area and a foreground tissue area and standardizing colors.
Dividing the pathological image block into a background blank area and a foreground tissue area to obtain small blocks which can be directly used for analyzing the foreground tissue area;
and after the small blocks to be processed are obtained, performing color conversion according to the color distribution of the training set samples to enhance the color variability of the data.
Step 4-1: and the pathological image subtype classification module analyzes the processed small blocks by using a deep learning model to obtain expected classification.
Medical image classification refers to a technique for classifying medical images into diagnoses of whether or not they are suffering from a certain disease or graded in severity. In this embodiment, image classification processing is performed on the digital pathological image blocks to obtain a specific subtype classification result.
Optionally, the pathological subtypes include: normal tissue area, tumor area. Tumors in different locations correspond to different subtype pathology classifications. For example, breast cancer pathotyping comprises: non-invasive cancer (intraductal carcinoma, lobular carcinoma in situ, intraductal papillary carcinoma, eczematoid breast carcinoma), early invasive cancer (early invasive ductal carcinoma, early invasive lobular carcinoma), invasive cancer (invasive specific cancer, invasive non-specific cancer), rare cancer. The invasive specific cancers include papillary carcinoma, medullary carcinoma (with large amount of lymphocyte infiltration), tubular carcinoma (high differentiation adenocarcinoma), adenoid cystic carcinoma, mucinous adenocarcinoma, apocrine adenoid carcinoma, and squamous cell carcinoma. The lung cancer pathological typing comprises: small cell lung cancer and non-small cell lung cancer (squamous cell lung cancer, adenocarcinoma, large cell carcinoma. among them, adenocarcinoma is further classified into adherent growth type, acinar growth type, papillary shape, microemulsion head shape, and solid growth type.
Alternatively, pathological image subtype classification can also be combined with gene expression; the correlation between the tumor tissue genotype mutation and the expression of the gene in the tissue and cells is established.
When the pathological image blocks are classified, a classic machine learning classification model or a deep learning classification model can be adopted. A classical machine learning classification model comprising: the system comprises a K nearest neighbor classification model, a Bayesian classification model, a random forest classification model, a support vector machine model and the like. The deep learning model can adopt one of classical classification models such as LeNet, AlexNet, VGG, GoogLeNet, ResNet and the like to perform image classification processing, and can also adopt other classification networks or designed specific network models. Alternatively, a pre-trained version of the classical model is used for the transfer learning. In the present disclosure, the pathological image is analyzed using the deep learning model, and when the pathological image block is classified, the specific classification model used is not limited.
Step 4-2: and the pathological image subtype segmentation module analyzes the processed small blocks by using a deep learning model to obtain expected segmentation.
Image segmentation is a technique and process that divides an image into several specific regions with unique properties and proposes an object of interest. Medical image segmentation is the identification and extraction of external contours and internal voxels of an object of interest in a medical image, enabling the segmentation of an organ or lesion region in the image. In this embodiment, the digital pathological image block is segmented to obtain a specific region of interest. Usually, the segmentation of cell nuclei in histopathological sections is realized, and the obtained segmentation of cell nuclei can be used for automatic counting and characteristic quantitative description of cells.
When pathological image blocks are segmented, a classical segmentation method or a deep learning segmentation model can be adopted. A classical segmentation method comprising: threshold-based segmentation methods, region-based segmentation methods, edge-based segmentation methods, histogram-based segmentation methods, and segmentation methods based on specific theories. The deep learning model can adopt one of FCN, U-Net + +, SegNet, RefineNet and other classical segmentation models to carry out image segmentation processing, and can also adopt other segmentation networks or designed specific network models. Alternatively, a pre-trained version of the classical model is used for the transfer learning. In the present disclosure, the pathological image is analyzed using the deep learning model, and when the pathological image block segmentation is realized, the specific segmentation model to be used is not limited.
Step 4-3: and the pathological image subtype detection module analyzes the processed small blocks by using a deep learning model to obtain expected detection.
Object detection is a computer vision image processing related technique for detecting semantic objects of a particular class in a digital image or video. Medical image object detection is commonly used to detect and identify specific medical objects. In this embodiment, the target in the digital pathological image block is detected to obtain a specific subtype detection result, or the cell nucleus in the image block is detected.
In one embodiment, nuclei in the pathological image block are detected for further analysis statistics. Morphological characteristics of cell nuclei and distribution of cell nuclei in tissues are important criteria for cancer diagnosis.
Optionally, the growth pattern subtypes for lung adenocarcinoma pathology images include: adherent growth, acinar growth, papillary, and parenchymal growth. The microemulsion head type has low differentiation degree, high malignancy degree, strong invasiveness and easy metastasis, and occupies an important position in the pathological diagnosis of lung adenocarcinoma. The method has the advantages that the microemulsion head type in the lung adenocarcinoma pathological image is detected, the automatic marking of the microemulsion head type is realized, and a pathologist can find the microemulsion head type tissue more easily, so that the diagnosis efficiency and accuracy are improved.
When the targets in the pathological image blocks are detected, a classical detection method or a deep learning model can be adopted. A classical detection method comprising: Viola-Jones detector, HOG detector, deformable component model, etc. The deep learning model can adopt one of typical target detection models such as R-CNN, SPPNet, Fast-RCNN, Mask-RCNN, YOLO, SSD, Retina-Net and the like to carry out image target detection processing, and can also adopt other detection models or designed specific network models. Alternatively, a pre-trained version of the classical model is used for the transfer learning. It should be noted that, in the present disclosure, the deep learning model is used to analyze the pathological image, and when the target detection of the pathological image block is implemented, the specific target detection model is not limited.
Further, as shown in fig. 2, a procedure for acquiring a full-slice digital pathology image is as follows:
step 1: encapsulating the excised pathological tissue in a wax block comprising: obtaining tissue specimen, fixing, washing, dehydrating, removing residual water and fat, and permeating wax.
Step 2: the waxed tissue is processed by embedding, slicing, sticking, dewaxing and rehydration, H & E dyeing, washing, differentiation, rinsing, dehydration, counterdyeing, dehydration, mounting and the like. Paraffin blocks were cut into very thin high quality sections (sections) which were mounted on glass slides and appropriately stained to reveal normal and abnormal structures.
And step 3: after obtaining the pathological section, the whole slide is scanned by using a microscope system to obtain a full-section digital pathological image. The pathological image capturing device is not limited in this embodiment.
Further, as shown in fig. 3, the full-slice digital pathology image preprocessing procedure includes:
step 1: and dividing the pathological image block into a background blank area and a foreground tissue area to obtain small blocks which can be directly used for analyzing the foreground tissue area.
The method for distinguishing the background blank area from the foreground tissue area may be a threshold-based method, a histogram statistical method, or other methods implemented by a network model.
Step 2: and performing color conversion to enhance the color variability of the data.
Histologically, a stain (often abbreviated as H & E) of a combination of hematoxylin and eosin is often used. Hematoxylin stains the nucleus blue and eosin stains the cell plasma red. The color of the section is related to the dyeing conditions such as solution concentration, processing time, temperature and the like during dyeing processing, and the color distribution presented by the dyed section is different, so that the appearance is obviously changed. In order to make a reliable diagnosis, it is necessary to correct color variations in the pathological image.
Where the variability in the appearance of histopathological images is accounted for by performing normalization of the images prior to analysis, color normalization is typically performed. Such methods effectively standardize color while maintaining structural integrity.
When the color normalization is performed on the pathological image block, a natural scene image color enhancement method or a color deconvolution algorithm can be adopted, or a countermeasure model based on the deep learning field can be adopted. A method of color enhancement of images of natural scenes, comprising: histogram equalization, histogram standardization, Retinex enhancement and the like, and the method does not consider the essential characteristics of pathological images and is difficult to obtain good normalization effect. Most effective pathological image normalization methods are based on a color deconvolution-based staining separation method, and algorithms perform linear transformation on pathological images in optical density space to separate independent color components. Such methods typically first separate the nucleus from the cytoplasmic region, which is used as a priori to calculate adaptive color deconvolution parameters. The method based on the deep learning domain confrontation ignores irrelevant appearance variability through the domain confrontation training, and can analyze the performance of a problem based on a single histopathology image to obtain a dyeing normalization result.
In the present disclosure, when color normalization of a pathological image block is implemented, a specific color normalization method is not limited.
Further, the label processing procedure corresponding to the full-slice pathological image includes:
step 1: and (5) labeling processing. And reading and labeling the target structure of the digital pathological image by using pathological image labeling software. The target structure includes: tumor region, non-tumor region subtype region, or different tissue student long pattern subtype region. It should be noted that the various subtypes do not overlap.
In particular, using the QuPath software to label subtype regions of a pathology image, region labeling can be achieved at different magnifications. Newly building a project and importing pictures in QuPath software; and defining a marking type and a corresponding color number. Note that the callout may not have overlapping regions.
Step 2: and deriving the label picture. The label image corresponding to the digital pathology image is derived from software by using a script supported by QuPath, and can be derived into a PNG image format, wherein the color of the image corresponds to the subtype of the pathology. Depending on the specific requirements and image size, it is possible to choose to derive 2-fold down-sampled or 4-fold down-sampled label images.
And step 3: and (4) blocking, namely blocking the label picture into small label blocks of the image corresponding to the digital pathological image according to the full-slice digital pathological image blocking processing rule.
And 4, step 4: the small blocks of the full-slice pathological image correspond to the small blocks of the label, and the classified, segmented or detected label corresponding to each pathological image block is obtained.
Example two:
the embodiment aims to provide an automatic analysis method of a full-slice digital pathological image.
A method for automatic analysis of a full-slice digital pathology image, comprising:
dividing the acquired full-slice digital pathological image into a plurality of image small blocks;
dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks;
performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and analyzing the processed image small blocks by utilizing a pre-trained deep learning model, and analyzing the full-slice digital pathological image according to an analysis processing result.
The method comprises the steps of data acquisition, slicing, preprocessing, label labeling, classification, segmentation and target detection by using the deep learning model. Including the training and testing phases of the model.
Example three:
the embodiment aims at providing an electronic device.
An electronic device comprising a memory, a processor and a computer program stored in the memory for execution by the processor, the program when executed by the processor implementing a method for automatic analysis of a full-slice digital pathology image, comprising:
dividing the acquired full-slice digital pathological image into a plurality of image small blocks;
dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks;
performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and analyzing the processed image small blocks by utilizing a pre-trained deep learning model, and analyzing the full-slice digital pathological image according to an analysis processing result.
Example four:
an object of the present embodiment is to provide a computer-readable storage medium.
A computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of automatic analysis of a full-slice digital pathology image, comprising:
dividing the acquired full-slice digital pathological image into a plurality of image small blocks;
dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks;
performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and analyzing the processed image small blocks by utilizing a pre-trained deep learning model, and analyzing the full-slice digital pathological image according to an analysis processing result.
Those skilled in the art will appreciate that the modules or steps of the present disclosure described above can be implemented using general purpose computer means, or alternatively, they can be implemented using program code executable by computing means, such that they are stored in a storage means and executed by computing means, or they are separately fabricated into individual integrated circuit modules, or multiple modules or steps thereof are fabricated into a single integrated circuit module. The present disclosure is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
Although the present disclosure has been described with reference to specific embodiments, it should be understood that the scope of the present disclosure is not limited thereto, and those skilled in the art will appreciate that various modifications and changes can be made without departing from the spirit and scope of the present disclosure.
Claims (10)
1. An automated analysis system for full-slice digital pathology images, comprising:
the image preprocessing module is used for dividing the acquired full-slice digital pathological image into a plurality of image small blocks; dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks; performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and the automatic analysis module is used for analyzing and processing the processed image small blocks by utilizing the pre-trained deep learning model and realizing the analysis of the full-slice pathological image according to the analysis and processing result.
2. The system for automatic analysis of full-slice digital pathology images of claim 1, wherein said automatic analysis module comprises:
the classification unit is used for predicting the image small blocks by using the trained deep learning model to obtain a prediction classification result of each pathological image small block;
the segmentation unit is used for predicting the image small blocks by using the trained deep learning model to obtain the prediction segmentation result of each pathological image small block;
and the target detection unit is used for predicting the image small blocks by using the trained deep learning model to obtain a prediction target detection result of each pathological image small block.
3. The system of claim 1, wherein the automatic analysis module marks the image patches according to the classification result, the segmentation result and the target detection result obtained by the analysis process, and splices the marked image patches back to the size of the original input image.
4. An automatic analysis system of full-slice digital pathology images according to claim 1, characterized in that the system further comprises a visualization unit for displaying marked pathology image patches and images stitched with marked case image patches.
5. An automatic analysis method for a full-slice digital pathology image, comprising:
dividing the acquired full-slice digital pathological image into a plurality of image small blocks;
dividing the pathological image small blocks into a background blank area and a foreground tissue area to obtain foreground tissue area image small blocks;
performing color conversion on the image small blocks according to the color distribution of the training set sample, and enhancing the color variability of data;
and analyzing the processed image small blocks by utilizing a pre-trained deep learning model, and analyzing the full-slice digital pathological image according to a processing result.
6. The method as claimed in claim 5, wherein the step of analyzing the full-slice digital pathological image according to the analysis processing result includes marking the prediction result of the automatic analysis module on each pathological image patch, and splicing the marked image patches back to the size of the original input image.
7. The method as claimed in claim 6, wherein the stitched image is displayed on a display, and the user can assist in determining the location and type of lesion in the image according to the labeling result.
8. The method of claim 5, wherein the analysis process comprises image classification, image segmentation and target detection of the image on the image patch.
9. An electronic device comprising a memory, a processor and a computer program stored and executed on the memory, wherein the processor when executing the program implements a method for automatic analysis of a full slice digital pathology image according to any one of claims 5-8.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out a method of automatic analysis of a full-slice digital pathology image according to any one of claims 5-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011346662.6A CN112435243A (en) | 2020-11-26 | 2020-11-26 | Automatic analysis system and method for full-slice digital pathological image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011346662.6A CN112435243A (en) | 2020-11-26 | 2020-11-26 | Automatic analysis system and method for full-slice digital pathological image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112435243A true CN112435243A (en) | 2021-03-02 |
Family
ID=74697796
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011346662.6A Pending CN112435243A (en) | 2020-11-26 | 2020-11-26 | Automatic analysis system and method for full-slice digital pathological image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112435243A (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112950599A (en) * | 2021-03-10 | 2021-06-11 | 中山大学 | Large intestine cavity area and intestine content labeling method based on deep learning |
| CN113222903A (en) * | 2021-04-20 | 2021-08-06 | 山东师范大学 | Full-section histopathology image analysis method and system |
| CN113313705A (en) * | 2021-06-22 | 2021-08-27 | 上海杏脉信息科技有限公司 | Pathological image processing system, method and medium |
| CN113344849A (en) * | 2021-04-25 | 2021-09-03 | 山东师范大学 | Microemulsion head detection system based on YOLOv5 |
| CN113469939A (en) * | 2021-05-26 | 2021-10-01 | 透彻影像(北京)科技有限公司 | HER-2 immunohistochemical automatic interpretation system based on characteristic curve |
| CN113762395A (en) * | 2021-09-09 | 2021-12-07 | 深圳大学 | Pancreatic bile duct type ampulla carcinoma classification model generation method and image classification method |
| CN114387201A (en) * | 2021-04-08 | 2022-04-22 | 透彻影像科技(南京)有限公司 | Cytopathic image auxiliary diagnosis system based on deep learning and reinforcement learning |
| US11348228B2 (en) | 2017-06-26 | 2022-05-31 | The Research Foundation For The State University Of New York | System, method, and computer-accessible medium for virtual pancreatography |
| CN114862798A (en) * | 2022-05-09 | 2022-08-05 | 华东师范大学 | Multi-view representation learning method for tumor pathology auxiliary diagnosis |
| CN114897823A (en) * | 2022-05-10 | 2022-08-12 | 广州锟元方青医疗科技有限公司 | Cytology sample image quality control method, system and storage medium |
| CN114937045A (en) * | 2022-06-20 | 2022-08-23 | 四川大学华西医院 | Hepatocellular carcinoma pathological image segmentation system |
| CN114943723A (en) * | 2022-06-08 | 2022-08-26 | 北京大学口腔医学院 | Method for segmenting and counting irregular cells and related equipment |
| CN114972341A (en) * | 2022-07-28 | 2022-08-30 | 华南理工大学 | WSI image classification method, system and medium based on Bayesian assisted learning |
| WO2022267379A1 (en) * | 2021-06-22 | 2022-12-29 | 上海杏脉信息科技有限公司 | Pathological image processing and display system, client, server and medium |
| TWI805290B (en) * | 2022-03-28 | 2023-06-11 | 臺北醫學大學 | Method for predicting whether lung adenocarcinoma has epidermal growth factor receptor mutations |
| CN116309585A (en) * | 2023-05-22 | 2023-06-23 | 山东大学 | Method and system for identifying breast ultrasound image target area based on multitask learning |
| CN116386902A (en) * | 2023-04-24 | 2023-07-04 | 北京透彻未来科技有限公司 | Artificial intelligent auxiliary pathological diagnosis system for colorectal cancer based on deep learning |
| CN116664550A (en) * | 2023-07-10 | 2023-08-29 | 广州医科大学附属第一医院（广州呼吸中心） | Intelligent identification method and device for lung cancer tissue immunohistochemical PD-L1 pathological section |
| CN117314888A (en) * | 2023-11-02 | 2023-12-29 | 中山大学附属第一医院 | Crohn disease detection method based on multi-example learning and pathological image |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550651A (en) * | 2015-12-14 | 2016-05-04 | 中国科学院深圳先进技术研究院 | Method and system for automatically analyzing panoramic image of digital pathological section |
| CN108172278A (en) * | 2018-01-24 | 2018-06-15 | 北京航空航天大学 | A kind of HE dyes pathological image color method for normalizing |
| CN108717554A (en) * | 2018-05-22 | 2018-10-30 | 复旦大学附属肿瘤医院 | A kind of thyroid tumors histopathologic slide image classification method and its device |
| CN111462076A (en) * | 2020-03-31 | 2020-07-28 | 湖南国科智瞳科技有限公司 | Method and system for detecting fuzzy area of full-slice digital pathological image |
| CN111554384A (en) * | 2020-05-14 | 2020-08-18 | 南京信息工程大学 | Adenocarcinoma pathological image analysis method based on prior perception and multitask learning |
| CN111784711A (en) * | 2020-07-08 | 2020-10-16 | 麦克奥迪（厦门）医疗诊断系统有限公司 | Lung pathology image classification and segmentation method based on deep learning |
-
2020
- 2020-11-26 CN CN202011346662.6A patent/CN112435243A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550651A (en) * | 2015-12-14 | 2016-05-04 | 中国科学院深圳先进技术研究院 | Method and system for automatically analyzing panoramic image of digital pathological section |
| CN108172278A (en) * | 2018-01-24 | 2018-06-15 | 北京航空航天大学 | A kind of HE dyes pathological image color method for normalizing |
| CN108717554A (en) * | 2018-05-22 | 2018-10-30 | 复旦大学附属肿瘤医院 | A kind of thyroid tumors histopathologic slide image classification method and its device |
| CN111462076A (en) * | 2020-03-31 | 2020-07-28 | 湖南国科智瞳科技有限公司 | Method and system for detecting fuzzy area of full-slice digital pathological image |
| CN111554384A (en) * | 2020-05-14 | 2020-08-18 | 南京信息工程大学 | Adenocarcinoma pathological image analysis method based on prior perception and multitask learning |
| CN111784711A (en) * | 2020-07-08 | 2020-10-16 | 麦克奥迪（厦门）医疗诊断系统有限公司 | Lung pathology image classification and segmentation method based on deep learning |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11348228B2 (en) | 2017-06-26 | 2022-05-31 | The Research Foundation For The State University Of New York | System, method, and computer-accessible medium for virtual pancreatography |
| CN112950599A (en) * | 2021-03-10 | 2021-06-11 | 中山大学 | Large intestine cavity area and intestine content labeling method based on deep learning |
| CN114387201A (en) * | 2021-04-08 | 2022-04-22 | 透彻影像科技(南京)有限公司 | Cytopathic image auxiliary diagnosis system based on deep learning and reinforcement learning |
| CN113222903A (en) * | 2021-04-20 | 2021-08-06 | 山东师范大学 | Full-section histopathology image analysis method and system |
| CN113344849B (en) * | 2021-04-25 | 2023-01-24 | 山东师范大学 | Microemulsion head detection system based on YOLOv5 |
| CN113344849A (en) * | 2021-04-25 | 2021-09-03 | 山东师范大学 | Microemulsion head detection system based on YOLOv5 |
| CN113469939A (en) * | 2021-05-26 | 2021-10-01 | 透彻影像(北京)科技有限公司 | HER-2 immunohistochemical automatic interpretation system based on characteristic curve |
| WO2022267379A1 (en) * | 2021-06-22 | 2022-12-29 | 上海杏脉信息科技有限公司 | Pathological image processing and display system, client, server and medium |
| CN113313705A (en) * | 2021-06-22 | 2021-08-27 | 上海杏脉信息科技有限公司 | Pathological image processing system, method and medium |
| CN113762395A (en) * | 2021-09-09 | 2021-12-07 | 深圳大学 | Pancreatic bile duct type ampulla carcinoma classification model generation method and image classification method |
| TWI805290B (en) * | 2022-03-28 | 2023-06-11 | 臺北醫學大學 | Method for predicting whether lung adenocarcinoma has epidermal growth factor receptor mutations |
| CN114862798A (en) * | 2022-05-09 | 2022-08-05 | 华东师范大学 | Multi-view representation learning method for tumor pathology auxiliary diagnosis |
| CN114897823A (en) * | 2022-05-10 | 2022-08-12 | 广州锟元方青医疗科技有限公司 | Cytology sample image quality control method, system and storage medium |
| CN114897823B (en) * | 2022-05-10 | 2024-03-19 | 广州锟元方青医疗科技有限公司 | Cytological sample image quality control method, system and storage medium |
| CN114943723A (en) * | 2022-06-08 | 2022-08-26 | 北京大学口腔医学院 | Method for segmenting and counting irregular cells and related equipment |
| CN114937045A (en) * | 2022-06-20 | 2022-08-23 | 四川大学华西医院 | Hepatocellular carcinoma pathological image segmentation system |
| CN114972341A (en) * | 2022-07-28 | 2022-08-30 | 华南理工大学 | WSI image classification method, system and medium based on Bayesian assisted learning |
| CN114972341B (en) * | 2022-07-28 | 2022-12-16 | 华南理工大学 | WSI image classification method, system and medium based on Bayesian assisted learning |
| CN116386902A (en) * | 2023-04-24 | 2023-07-04 | 北京透彻未来科技有限公司 | Artificial intelligent auxiliary pathological diagnosis system for colorectal cancer based on deep learning |
| CN116386902B (en) * | 2023-04-24 | 2023-12-19 | 北京透彻未来科技有限公司 | Artificial intelligent auxiliary pathological diagnosis system for colorectal cancer based on deep learning |
| CN116309585A (en) * | 2023-05-22 | 2023-06-23 | 山东大学 | Method and system for identifying breast ultrasound image target area based on multitask learning |
| CN116309585B (en) * | 2023-05-22 | 2023-08-22 | 山东大学 | Method and system for identifying breast ultrasound image target area based on multitask learning |
| CN116664550A (en) * | 2023-07-10 | 2023-08-29 | 广州医科大学附属第一医院（广州呼吸中心） | Intelligent identification method and device for lung cancer tissue immunohistochemical PD-L1 pathological section |
| CN116664550B (en) * | 2023-07-10 | 2024-04-12 | 广州医科大学附属第一医院（广州呼吸中心） | Intelligent identification method and device for lung cancer tissue immunohistochemical PD-L1 pathological section |
| CN117314888A (en) * | 2023-11-02 | 2023-12-29 | 中山大学附属第一医院 | Crohn disease detection method based on multi-example learning and pathological image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112435243A (en) | Automatic analysis system and method for full-slice digital pathological image | |
| EP3486836B1 (en) | Image analysis method, apparatus, program, and learned deep learning algorithm | |
| US10650221B2 (en) | Systems and methods for comprehensive multi-assay tissue analysis | |
| CN111448569B (en) | Method for storing and retrieving digital pathology analysis results | |
| JP2021506022A (en) | Deep learning systems and methods for solidarity cell and region classification in bioimaging | |
| CN113454733A (en) | Multi-instance learner for prognostic tissue pattern recognition | |
| CN107430771A (en) | System And Method For Image Segmentation | |
| US11538261B2 (en) | Systems and methods for automated cell segmentation and labeling in immunofluorescence microscopy | |
| US11176412B2 (en) | Systems and methods for encoding image features of high-resolution digital images of biological specimens | |
| US11959848B2 (en) | Method of storing and retrieving digital pathology analysis results | |
| CN112990214A (en) | Medical image feature recognition prediction model | |
| CN113261012B (en) | Method, device and system for processing image | |
| CN113393454A (en) | Method and device for segmenting pathological target examples in biopsy tissues | |
| US11887355B2 (en) | System and method for analysis of microscopic image data and for generating an annotated data set for classifier training | |
| US20230162485A1 (en) | Digital analysis of preanalytical factors in tissues used for histological staining | |
| Sabata et al. | Automated analysis of pin-4 stained prostate needle biopsies | |
| US20230410316A1 (en) | Sequential convolutional neural networks for nuclei segmentation | |
| Selcuk et al. | Automated HER2 Scoring in Breast Cancer Images Using Deep Learning and Pyramid Sampling | |
| Botega et al. | Quantification of histological neoplastic cells using digital image processing | |
| US20210366107A1 (en) | Sensitivity analysis for digital pathology | |
| Lor et al. | FAST INTERACTIVE REGIONAL PATTERN MERGING FOR GENERIC TISSUE SEGMENTATION IN HISTOPATHOLOGY IMAGES | |
| KR20240012738A (en) | Cluster analysis system and method of artificial intelligence classification for cell nuclei of prostate cancer tissue | |
| Priya | Histopathological slide analysis using digital microscopy | |
| Blackledge et al. | Pattern Recognition in Cytopathology for Papanicolaou Screening | |
| Sahu | Introductory Review on Computational Histopathology |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210302 |