CN116580397B - Pathological image recognition method, device, equipment and storage medium - Google Patents
Pathological image recognition method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN116580397B CN116580397B CN202310851588.0A CN202310851588A CN116580397B CN 116580397 B CN116580397 B CN 116580397B CN 202310851588 A CN202310851588 A CN 202310851588A CN 116580397 B CN116580397 B CN 116580397B
- Authority
- CN
- China
- Prior art keywords
- pathological
- region
- image
- feature
- candidate
- 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
- 230000001575 pathological effect Effects 0.000 title claims abstract description 293
- 238000000034 method Methods 0.000 title claims abstract description 80
- 238000012545 processing Methods 0.000 claims abstract description 112
- 238000013507 mapping Methods 0.000 claims abstract description 28
- 238000010586 diagram Methods 0.000 claims abstract description 27
- 238000007781 pre-processing Methods 0.000 claims abstract description 20
- 230000007170 pathology Effects 0.000 claims description 62
- 238000012790 confirmation Methods 0.000 claims description 14
- 238000012706 support-vector machine Methods 0.000 claims description 13
- 238000001514 detection method Methods 0.000 claims description 12
- 238000011176 pooling Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 7
- 238000005728 strengthening Methods 0.000 claims description 7
- 238000012935 Averaging Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 238000003708 edge detection Methods 0.000 claims description 5
- 239000003086 colorant Substances 0.000 claims description 4
- 238000012847 principal component analysis method Methods 0.000 claims description 4
- 238000003745 diagnosis Methods 0.000 abstract description 10
- 210000004027 cell Anatomy 0.000 description 18
- 239000013598 vector Substances 0.000 description 15
- 230000008569 process Effects 0.000 description 8
- 238000004891 communication Methods 0.000 description 6
- 230000000877 morphologic effect Effects 0.000 description 5
- 230000011218 segmentation Effects 0.000 description 5
- 238000010191 image analysis Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 210000004881 tumor cell Anatomy 0.000 description 2
- 241001270131 Agaricus moelleri Species 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000010827 pathological analysis Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000010186 staining Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/695—Preprocessing, e.g. image segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- 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
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- 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
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
-
- 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
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/698—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of computer medical application, and discloses a pathological image identification method, a device, equipment and a storage medium, wherein the method comprises the following steps: carrying out gray level processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray level processing to obtain a binary image; denoising the binary image to generate a preprocessing pathological image; extracting features of the preprocessed pathological image to obtain a feature area diagram; identifying whether the target pathological features exist in the feature area diagram, and confirming the area with the target pathological features as a pathological area; mapping the pathological region to a characteristic region map to obtain corresponding characteristics of the pathological region; classifying the corresponding features of the pathological region to obtain feature categories; acquiring positioning data of a pathological area; and generating a recognition result of the pathological image according to the characteristic category and the positioning data. According to the invention, the diagnosis efficiency and accuracy of a pathologist can be further improved by carrying out feature recognition, classification and positioning on the preprocessed pathological images.
Description
Technical Field
The present invention relates to the field of computer medical applications, and in particular, to a method, an apparatus, a device, and a storage medium for identifying pathological images.
Background
Pathological images play a vital role in medical diagnostics. At present, the digital processing of pathological images becomes a hot research direction in the medical field. The digital pathological image technology can further improve the diagnosis efficiency and accuracy of doctors. And pathology image technology is gradually involved in multiple fields of computer vision, image processing, artificial intelligence and the like.
However, the traditional manual identification process by a pathologist faces many challenges, and the manual film reading process is tedious and time-consuming. Thus, extracting disease-related features from pathology images is a challenging problem. Therefore, a pathological image recognition method is required to improve the diagnosis efficiency and accuracy of pathologists.
The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present invention and is not intended to represent an admission that the foregoing is prior art.
Disclosure of Invention
The invention mainly aims to provide a pathological image recognition method, a device, equipment and a storage medium, and aims to solve the technical problem that relevant pathological features of pathological images are difficult to accurately recognize in the prior art.
To achieve the above object, the present invention provides a pathological image recognition method comprising the steps of:
Carrying out gray level processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray level processing by a maximum inter-class variance method to obtain a binary image;
denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image;
extracting features of the preprocessed pathological image based on a single-step multi-frame target detection model to obtain a feature area diagram;
identifying whether a target pathological feature exists in the feature area diagram through an edge frame network, and confirming an area with the target pathological feature as a pathological area;
mapping the pathological region to the characteristic region map through a region suggestion pooling back propagation algorithm to obtain a pathological region corresponding characteristic;
classifying the corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories;
acquiring positioning data of the pathological area based on a regression algorithm;
and generating a recognition result of the pathological image according to the characteristic category and the positioning data.
Optionally, the step of identifying whether the target pathological feature exists in the feature area map through an edge frame network and identifying the area where the target pathological feature exists as a pathological area includes:
Dividing the characteristic area map into a plurality of first candidate area maps;
identifying whether target pathological features exist in the first candidate region map through an edge frame network;
confirming a first candidate region map with the target pathological features as a candidate pathological region;
dividing the candidate pathological area into a plurality of second candidate area maps;
respectively acquiring Jaccard similarity between corresponding features of a plurality of second candidate region graphs and the target pathological features;
and confirming the second candidate region map with the Jaccard similarity higher than a preset similarity threshold as a pathological region.
Optionally, the step of dividing the candidate pathological area into a plurality of second candidate area maps includes:
detecting a cell state within the candidate pathological region;
labeling the foreground and the background of the region of the attached cell in the candidate pathological region, and dividing the labeled attached cell in the region of the attached cell based on a watershed algorithm to obtain a first intermediate processing image;
performing division-free preprocessing on the complete features in the first intermediate processing image through an edge detection algorithm to obtain a second intermediate processing image;
the second intermediate processed image is divided into a plurality of second candidate region maps based on a region growing algorithm.
Optionally, the step of performing gray processing on the pathological image, performing binarization processing on the pathological image after gray processing by using a maximum inter-class variance method, and obtaining a binary image includes:
adjusting the primary colors of the pathological image according to a weighted average method to realize gray scale processing;
detecting the number of pixels corresponding to each gray level in the pathological image after gray level processing to obtain a gray level mapping table;
traversing the gray mapping table based on a maximum inter-class variance method to generate a target gray threshold;
and performing binarization processing on the pathological image subjected to gray level processing according to the target gray level threshold value to obtain a binary image.
Optionally, the step of traversing the gray mapping table based on the maximum inter-class variance method to generate the target gray threshold value includes:
traversing the gray mapping table based on a maximum inter-class variance method to obtain a gray traversing result;
determining an image processing type according to the pathology information corresponding to the pathology image;
confirming corresponding operation of the image processing type through an image processing relation table;
when the image processing type corresponds to the cell edge characteristic strengthening operation, the Matlab platform is combined with the gray level traversing result to adjust the gray level threshold value of the image, so that the cell edge characteristic strengthening operation is completed, and the current gray level threshold value is obtained;
And taking the current gray threshold value as a target gray threshold value.
Optionally, the step of classifying the corresponding features of the pathological area based on the support vector machine algorithm to obtain feature categories includes:
detecting the feature dimension level of the corresponding feature of the pathological region;
when the feature dimension level is higher than a preset feature dimension level, performing feature dimension reduction processing on the corresponding features of the pathological region through a principal component analysis method;
classifying the processed corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories.
Optionally, the step of denoising the binary image by a neighborhood averaging algorithm to generate a preprocessed pathological image includes:
generating an image processing type according to the pathology information corresponding to the pathology image;
determining a neighborhood pixel size and a neighborhood comparison threshold according to the image processing type;
denoising the binary image according to the neighborhood pixel size and the neighborhood comparison threshold value through a neighborhood average algorithm to generate a preprocessing pathological image.
In addition, in order to achieve the above object, the present invention also proposes a pathological image recognition apparatus, the apparatus comprising:
The gray processing module is used for carrying out gray processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray processing through a maximum inter-class variance method to obtain a binary image;
the image denoising module is used for denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image:
the feature extraction module is used for carrying out feature extraction on the preprocessing pathological image based on the single-step multi-frame target detection model to obtain a feature area diagram;
the region confirmation module is used for recognizing whether the target pathological features exist in the feature region diagram through an edge frame network and confirming the region with the target pathological features as a pathological region;
the feature mapping module is used for mapping the pathological region to the feature region map through a region suggestion pooling back propagation algorithm to obtain the corresponding feature of the pathological region;
the feature classification module is used for classifying the corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories;
the pathology positioning module is used for acquiring positioning data of the pathology area based on a regression algorithm;
and the result generation module is used for generating a recognition result of the pathological image according to the characteristic category and the positioning data.
In addition, in order to achieve the above object, the present invention also proposes a pathology image recognition apparatus, the apparatus comprising: a memory, a processor and a pathology image recognition program stored on the memory and executable on the processor, the pathology image recognition program being configured to implement the steps of the pathology image recognition method as described above.
In addition, in order to achieve the above object, the present invention also proposes a storage medium having stored thereon a pathology image recognition program which, when executed by a processor, implements the steps of the pathology image recognition method as described above.
The invention discloses a pathological image identification method, a pathological image identification device, pathological image identification equipment and a storage medium, wherein the pathological image identification method comprises the following steps: carrying out gray level processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray level processing by a maximum inter-class variance method to obtain a binary image; denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image; extracting features of the preprocessed pathological image to obtain a feature area diagram; identifying whether the target pathological features exist in the feature area diagram, and confirming the area with the target pathological features as a pathological area; mapping the pathological region to a characteristic region map to obtain corresponding characteristics of the pathological region; classifying the corresponding features of the pathological region to obtain feature categories; acquiring positioning data of a pathological area; and generating a recognition result of the pathological image according to the characteristic category and the positioning data. According to the invention, the diagnosis efficiency and accuracy of a pathologist can be further improved by carrying out feature recognition, classification and positioning on the preprocessed pathological images.
Drawings
FIG. 1 is a schematic structural diagram of a pathological image recognition device of a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a flowchart of a first embodiment of a pathological image recognition method according to the present invention;
FIG. 3 is a flowchart of a second embodiment of a pathological image recognition method according to the present invention;
FIG. 4 is a flowchart of a third embodiment of a pathological image recognition method according to the present invention;
fig. 5 is a block diagram showing a first embodiment of a pathological image recognition apparatus according to the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a pathological image recognition device of a hardware running environment according to an embodiment of the present invention.
As shown in fig. 1, the pathology image recognition apparatus may include: a processor 1001, such as a central processing unit (Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a Wireless interface (e.g., a Wireless-Fidelity (WI-FI) interface). The Memory 1005 may be a high-speed random access Memory (Random Access Memory, RAM) or a stable nonvolatile Memory (NVM), such as a disk Memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
It will be appreciated by those skilled in the art that the structure shown in fig. 1 does not constitute a limitation of the pathology image recognition apparatus, and may include more or fewer components than shown, or may combine certain components, or a different arrangement of components.
As shown in fig. 1, an operating system, a data storage module, a network communication module, a user interface module, and a pathology image recognition program may be included in the memory 1005 as one type of storage medium.
In the pathological image recognition apparatus shown in fig. 1, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the pathology image recognition apparatus of the present invention may be provided in the pathology image recognition apparatus, which invokes the pathology image recognition program stored in the memory 1005 through the processor 1001, and performs the pathology image recognition method provided by the embodiment of the present invention.
An embodiment of the invention provides a pathological image recognition method, referring to fig. 2, fig. 2 is a schematic flow chart of a first embodiment of the pathological image recognition method of the invention.
In this embodiment, the pathological image recognition method includes the following steps:
Step S10: and carrying out gray level processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray level processing by a maximum inter-class variance method to obtain a binary image.
It should be noted that, the execution subject of the method of the present embodiment may be an image recognition apparatus having functions of data processing, network communication, and program running, such as a pathology image recognition apparatus; other electronic devices having the same or similar functions or an image recognition system on which the electronic device is mounted may be used. The present embodiment and the following embodiments will exemplify a pathology image recognition method of the present invention with a pathology image recognition apparatus as an execution subject.
It is understood that the pathology image may be a pathology image of the pathology feature to be identified. The method for gray scale processing of the pathological image can be that the RGB value of each pixel is respectively averaged by an average value method to obtain a gray scale average value; a certain component (such as a red component) in the RGB value of each pixel may be used as a gray value by a component method; or converting the RGB value of each pixel into YUV color space by brightness method, and taking Y component as gray value; it is also possible to use a weighted average method: and carrying out weighted average on the RGB value of each pixel according to a certain weight value to obtain a gray value.
It should be understood that the maximum inter-class variance method (OTSU algorithm) may be an image segmentation algorithm, which is to divide an image into two parts by using gray characteristics of the image, and to determine a threshold value by calculating the variance of gray values of the image such that the inter-class variance between the foreground and the background is maximized under the threshold value, thereby segmenting the image into two parts.
Step S20: denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image.
It can be understood that the neighborhood averaging algorithm can be an image processing method, and the neighborhood averaging algorithm calculates an average value of pixels in a certain range around each pixel based on each pixel as a center, and replaces the value of the current pixel with the average value to achieve the effects of smoothing and denoising.
Further, when denoising the binary image, in order to achieve a better denoising effect, step S20 includes: generating an image processing type according to the pathology information corresponding to the pathology image; determining a neighborhood pixel size and a neighborhood comparison threshold according to the image processing type; denoising the binary image according to the neighborhood pixel size and the neighborhood comparison threshold value through a neighborhood average algorithm to generate a preprocessing pathological image.
It should be noted that, the pathology information corresponding to the pathology image may be related information of an inspection item corresponding to the pathology image, and an image processing type is generated according to the pathology information corresponding to the pathology image. For example: the pathological image is a tumor cell pathological image, and the tumor cells have irregular shapes and sizes, so that morphological influence factors are denoised, and the morphological influence factors are denoised according to the image processing types generated by the information.
It should be understood that the pathological image recognition device determines a neighborhood by using an image pixel as a center through a neighborhood averaging algorithm according to the image processing type, averages the pixel gray scales in the neighborhood, and obtains a target gray scale value; replacing the gray value of the current pixel with a target gray value; and carrying out the processing of the steps on each pixel to obtain the denoised pretreatment pathological image.
Step S30: and extracting the characteristics of the preprocessed pathological image based on the single-step multi-frame target detection model to obtain a characteristic area diagram.
It is understood that the single-step multi-frame object detection model (Single Shot MultiBox Detector, SSD) is a deep learning-based object detection model that can be used for feature extraction of pathological images and detection of pathological areas.
It should be understood that the pathological image recognition device may train the SSD model using the labeled pathological region as training data; inputting the preprocessed pathological image into a trained SSD model, extracting global characteristic information in the pathological image, and detecting a region possibly having pathological characteristics in the preprocessed pathological image, and the position and the size of the region by comparing and analyzing the characteristic information; marking the region where pathological features possibly exist in the preprocessing pathological image to generate a feature area diagram.
Step S40: and identifying whether the target pathological features exist in the feature area diagram through a border frame network, and confirming the area where the target pathological features exist as a pathological area.
It should be noted that the Edge Boxes network may be a fast object detection network, which may generate candidate Boxes in an image and then detect a target object by classifying the candidate Boxes. In the pathological image recognition process, the Edge Boxes network can be used for recognizing whether the target pathological features exist in the feature area diagram.
Step S50: and mapping the pathological region to the characteristic region map through a region suggestion pooling back propagation algorithm to obtain the corresponding characteristics of the pathological region.
It should be noted that, the region suggestion pooling back propagation algorithm (RoI alignment) may be a region interest pooling algorithm for target detection, which may map a pathological region onto a feature region map and obtain a feature representation corresponding to the pathological region.
It should be understood that the pathological image recognition apparatus maps the pathological region onto the characteristic region map according to the position and the size of the pathological region in the characteristic region map, and performs a pooling operation to extract the features corresponding to the pathological region from the characteristic region map.
Specifically, the pathological area is divided into a plurality of sub-areas, and bilinear interpolation operation is carried out on each sub-area through the RoI Align algorithm to obtain corresponding feature vectors. And splicing the feature vectors corresponding to the plurality of sub-regions to obtain the corresponding features of the pathological region.
Step S60: and classifying the corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories.
Further, when the corresponding features of the pathological region are classified by the support vector machine algorithm (Support Vector Machine, SVM), the operation efficiency of the algorithm may be affected due to the fact that feature dimensions are too high. Thus, step S60 further includes: detecting the feature dimension level of the corresponding feature of the pathological region; when the feature dimension level is higher than a preset feature dimension level, performing feature dimension reduction processing on the corresponding features of the pathological region through a principal component analysis method; classifying the processed corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories.
It should be noted that the feature dimension level may be the number of dimensions of the corresponding feature vector. Data is typically represented in the form of feature vectors, the number of dimensions of which represents the number of features, where each feature represents an attribute or feature of the data. And extracting feature vectors of the pathological region for feature detection tasks in pathological image recognition, wherein the number of dimensions of the feature vectors is the feature dimension level.
It should be understood that the preset feature dimension level may be a preset feature dimension level threshold, and when the feature dimension level threshold is exceeded, there may be too many non-pathological features, so that feature dimension reduction processing needs to be performed on the feature dimension level of the feature corresponding to the pathological region.
It is understood that the principal component analysis method may be an unsupervised dimension reduction method, in which high-dimensional feature vectors may be mapped into a low-dimensional space, and as much data variance as possible is preserved, thereby reducing the dimension of the feature vectors. And during specific treatment, the corresponding features of the pathological region are subjected to centering treatment, namely, the average value of each feature vector is subtracted, so that the average value is 0. And calculating a covariance matrix of the feature vector according to the centralized data. And carrying out eigenvalue decomposition on the covariance matrix to obtain eigenvalues and eigenvectors. And selecting the first k principal components according to the magnitude of the characteristic value, wherein the characteristic vectors corresponding to the k principal components are the basis vectors of the new characteristic space. And projecting the original data feature vector into a new feature space to obtain the feature vector after dimension reduction.
Step S70: and acquiring positioning data of the pathological region based on a regression algorithm.
It should be understood that the regression algorithm can predict the position and size information of the pathological region according to the input feature vector, and has higher positioning accuracy and robustness. In pathological image analysis and diagnosis, the positioning of pathological areas is a very important step, and accurate references can be provided for subsequent pathological diagnosis and treatment.
Step S80: and generating a recognition result of the pathological image according to the characteristic category and the positioning data.
It will be appreciated that the recognition result of the pathology image includes the classification type of the pathology feature and the location where the pathology feature is located. Diagnosis can be carried out through the identification result of the pathological image, so that the diagnosis efficiency and accuracy can be further improved.
In the embodiment, gray processing is performed on the pathological image, and binarization processing is performed on the pathological image after gray processing by a maximum inter-class variance method to obtain a binary image; denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image; extracting features of the preprocessed pathological image to obtain a feature area diagram; identifying whether the target pathological features exist in the feature area diagram, and confirming the area with the target pathological features as a pathological area; mapping the pathological region to a characteristic region map to obtain corresponding characteristics of the pathological region; classifying the corresponding features of the pathological region to obtain feature categories; acquiring positioning data of a pathological area; and generating a recognition result of the pathological image according to the characteristic category and the positioning data. According to the invention, the diagnosis efficiency and accuracy of a pathologist can be further improved by carrying out feature recognition, classification and positioning on the preprocessed pathological images.
Referring to fig. 3, fig. 3 is a flowchart illustrating a second embodiment of a pathological image recognition method according to the present invention.
Based on the first embodiment, in this embodiment, the step S40 includes:
step S401: the feature region map is divided into a plurality of first candidate region maps.
It should be noted that the purpose of dividing the feature region map into a plurality of first candidate region maps is to improve the efficiency and accuracy of pathological image analysis. In pathological image analysis, the order of magnitude of the feature region map is generally large, and the distribution of features is also relatively scattered. If the whole feature area diagram is directly analyzed, the calculation complexity is possibly high, and for the feature area diagram with a large size, a large amount of time and calculation resources are required to be consumed for processing, so that the processing efficiency is low. The feature distribution is not uniform. The feature distribution in the feature area graph is more scattered, and there may be a larger contribution of the features of some areas to the pathology analysis task, while the features of other areas have a smaller contribution to the task.
It can be appreciated that the feature region map is divided into a plurality of first candidate region maps, and each region is independently analyzed and processed in a block manner. Therefore, the calculation complexity can be effectively reduced, and the processing efficiency is improved. Meanwhile, by partitioning the features of different areas, the distribution situation of the features can be captured more accurately, and the accuracy of pathological image analysis is improved.
Step S402: and identifying whether the target pathological features exist in the first candidate region map through an edge frame network.
It should be noted that the step of identifying, through the Edge frame network, whether the target pathological feature exists in the first candidate region map may be that candidate frames are generated in the image through the Edge Boxes network, and then, whether the target pathological feature exists in the candidate region map is detected through classifying the candidate frames.
Step S403: and confirming the first candidate region map with the target pathological feature as a candidate pathological region.
Step S404: dividing the candidate pathological area into a plurality of second candidate area maps.
It should be understood that in performing similarity comparison between features, features in a pathological region need to be analyzed more finely, and then the pathological region needs to be divided into a second candidate region map that is more fine.
Further, in the process of dividing the second candidate region map, in order to make the features in the second candidate region map more obvious, the accuracy of the subsequent feature recognition is improved, and step S404 includes: detecting a cell state within the candidate pathological region; labeling the foreground and the background of the region of the attached cell in the candidate pathological region, and dividing the labeled attached cell in the region of the attached cell based on a watershed algorithm to obtain a first intermediate processing image; performing division-free preprocessing on the complete features in the first intermediate processing image through an edge detection algorithm to obtain a second intermediate processing image; the second intermediate processed image is divided into a plurality of second candidate region maps based on a region growing algorithm.
It will be appreciated that labeling the foreground and background can reduce errors and uncertainty by indicating which regions of the algorithm should be divided into foreground and which regions should be divided into background to assist the watershed algorithm in more accurate segmentation. The watershed algorithm determines a segmentation boundary according to gray values and gradient information in an image during segmentation. The labeling of the foreground and background can more accurately transfer the information to the algorithm, thereby enabling the segmentation boundary to be clearer and more accurate. Under the condition of no labeling information, the watershed algorithm can easily divide a small region in the image into isolated parts, so that the phenomenon of over-division is generated. By labeling the foreground and the background, the occurrence of the situation can be avoided, and the segmentation result is more reasonable and natural.
It will be appreciated that in dividing the pathological region into finer second candidate region maps, the complete features may be segmented due to the smaller area of the region, wherein the complete features may be complete morphological features including morphological structure, size, shape, arrangement of cells and tissues. Can be a nuclear feature: including the size, shape, staining properties of the nucleus, size, number, morphology, etc. The nuclear features can be used to distinguish between different kinds of cells and diseases. Texture features may also be: including texture information in the image such as texture density, granularity, directionality, etc. Texture features can be used to distinguish between different types of tissue and disease. Optical density characteristics may also be: including optical density values for different regions in the image, can be used to analyze tissue density and morphology. For example, the optical density of tumor tissue is generally higher than that of normal tissue.
In the specific implementation, the pathological image recognition device performs division-free preprocessing on the complete features in the first intermediate processing image through an edge detection algorithm to obtain a second intermediate processing image; the second intermediate processed image is then divided into a plurality of second candidate region maps based on a region growing algorithm.
Step S405: and respectively acquiring Jaccard similarity between the corresponding features of the second candidate region graphs and the target pathological features.
Step S406: and confirming the second candidate region map with the Jaccard similarity higher than a preset similarity threshold as a pathological region.
It is understood that the second candidate region map has finer complete pathological features, so that the Jaccard similarity between the corresponding features of the second candidate region map and the target pathological features can be obtained, and the target pathological features can be further accurately screened out. The screened area is identified as a pathological area.
In this embodiment, the feature area map is divided into a plurality of first candidate area maps; identifying whether target pathological features exist in the first candidate region map through an edge frame network; confirming a first candidate region map with the target pathological features as a candidate pathological region; dividing the candidate pathological area into a plurality of second candidate area maps; respectively acquiring Jaccard similarity between corresponding features of a plurality of second candidate region graphs and the target pathological features; and confirming the second candidate region map with the Jaccard similarity higher than a preset similarity threshold as a pathological region. According to the embodiment, the first candidate region map with the target pathological features is confirmed to be the candidate pathological region, and then the Jaccard similarity is obtained for the corresponding features of the second candidate region map and the target pathological features, so that more accurate pathological regions can be obtained, and the diagnosis efficiency and accuracy of pathologists are further improved.
Referring to fig. 4, fig. 4 is a flowchart illustrating a third embodiment of a pathological image recognition method according to the present invention.
Based on the above embodiments, in this embodiment, the step S10 includes:
step S101: and adjusting the primary colors of the pathological image according to a weighted average method to realize gray scale processing.
Step S102: and detecting the number of pixels corresponding to each gray level in the pathological image after gray level processing to obtain a gray level mapping table.
It can be understood that the pixel count can be realized through the image processing library by detecting the pixel number corresponding to each gray level in the pathological image after gray level processing. The ratio of the number of pixels per gray level to the total number of pixels is calculated and stored in the gray mapping table. The gray map may be a one-dimensional array, with each element in the gray map representing a gray level and the value of the element representing the ratio of the number of pixels of the gray level to the total number of pixels.
Step S103: and traversing the gray mapping table based on a maximum inter-class variance method to generate a target gray threshold.
Further, in order to generate a target gray threshold value more conforming to the target pathological feature corresponding to the pathological image, step S103 includes: traversing the gray mapping table based on a maximum inter-class variance method to obtain a gray traversing result; determining an image processing type according to the pathology information corresponding to the pathology image; confirming corresponding operation of the image processing type through an image processing relation table; when the image processing type corresponds to the cell edge characteristic strengthening operation, the Matlab platform is combined with the gray level traversing result to adjust the gray level threshold value of the image, so that the cell edge characteristic strengthening operation is completed, and the current gray level threshold value is obtained; and taking the current gray threshold value as a target gray threshold value.
It should be understood that, when the image processing type corresponds to the cell edge feature enhancement operation, the gray mapping table may be traversed for the cell edge feature in the pathological image, so as to obtain the number distribution situation of the pixels with different gray levels, so as to determine an appropriate gray threshold. And (3) realizing an image processing algorithm based on the Matlab platform, and performing cell edge characteristic strengthening operation on the pathological image according to the gray threshold value so as to highlight the outline and morphological characteristics of the cell edge. And (3) performing gray level traversal again according to the reinforced image to acquire a current gray level threshold value so as to verify the reinforcing effect and determine a final target gray level threshold value.
Step S104: and performing binarization processing on the pathological image subjected to gray level processing according to the target gray level threshold value to obtain a binary image.
In the specific implementation, the pathological image recognition device carries out binarization processing on the pathological image after gray processing according to a target gray threshold value generated based on a maximum inter-class variance method, and a corresponding binary image is obtained.
In the embodiment, the primary colors of the pathological image are adjusted according to a weighted average method, so that gray scale processing is realized; detecting the number of pixels corresponding to each gray level in the pathological image after gray level processing to obtain a gray level mapping table; traversing the gray mapping table based on a maximum inter-class variance method to generate a target gray threshold; and performing binarization processing on the pathological image subjected to gray level processing according to the target gray level threshold value to obtain a binary image. According to the method, the target gray threshold is generated through the maximum inter-class variance method, and then pathological image processing is carried out according to the target gray threshold, so that a binary image with better and clearer characteristics can be obtained.
Furthermore, an embodiment of the present invention proposes a storage medium having stored thereon a pathology image identification program which, when executed by a processor, implements the steps of the pathology image identification method as described above.
Referring to fig. 5, fig. 5 is a block diagram showing the structure of a first embodiment of a pathological image recognition apparatus according to the present invention.
As shown in fig. 5, a pathological image recognition device according to an embodiment of the present invention includes: a gray processing module 501, an image denoising module 502, a feature extraction module 503, a region confirmation module 504, a feature mapping module 505, a feature classification module 506, a pathology localization module 507, and a result generation module 508.
The gray processing module 501 is configured to perform gray processing on the pathological image, and perform binarization processing on the pathological image after gray processing by using a maximum inter-class variance method, so as to obtain a binary image.
The image denoising module 502 is configured to denoise the binary image by using a neighborhood averaging algorithm, and generate a preprocessed pathological image.
The feature extraction module 503 is configured to perform feature extraction on the preprocessed pathological image based on a single-step multi-frame target detection model, so as to obtain a feature area map.
The region confirmation module 504 is configured to identify, through a border frame network, whether a target pathological feature exists in the feature region map, and confirm a region in which the target pathological feature exists as a pathological region.
The feature mapping module 505 is configured to map the pathological region to the feature region map through a region suggestion pooling back propagation algorithm, so as to obtain a feature corresponding to the pathological region.
The feature classification module 506 is configured to classify the feature corresponding to the pathological area based on a support vector machine algorithm, and obtain a feature class.
The pathology locating module 507 is configured to obtain location data of the pathology region based on a regression algorithm.
The result generating module 508 is configured to generate a recognition result of the pathological image according to the feature class and the positioning data.
In the embodiment, gray processing is performed on the pathological image, and binarization processing is performed on the pathological image after gray processing by a maximum inter-class variance method to obtain a binary image; denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image; extracting features of the preprocessed pathological image to obtain a feature area diagram; identifying whether the target pathological features exist in the feature area diagram, and confirming the area with the target pathological features as a pathological area; mapping the pathological region to a characteristic region map to obtain corresponding characteristics of the pathological region; classifying the corresponding features of the pathological region to obtain feature categories; acquiring positioning data of a pathological area; and generating a recognition result of the pathological image according to the characteristic category and the positioning data. According to the invention, the diagnosis efficiency and accuracy of a pathologist can be further improved by carrying out feature recognition, classification and positioning on the preprocessed pathological images.
Other embodiments or specific implementation manners of the pathological image recognition device of the present invention may refer to the above method embodiments, and are not described herein.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. read-only memory/random-access memory, magnetic disk, optical disk), comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (8)
1. A pathology image recognition method, characterized in that the pathology image recognition method comprises the following steps:
carrying out gray level processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray level processing by a maximum inter-class variance method to obtain a binary image;
denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image;
extracting features of the preprocessed pathological image based on a single-step multi-frame target detection model to obtain a feature area diagram;
identifying whether a target pathological feature exists in the feature area diagram through an edge frame network, and confirming an area with the target pathological feature as a pathological area;
mapping the pathological region to the characteristic region map through a region suggestion pooling back propagation algorithm to obtain a pathological region corresponding characteristic;
classifying the corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories;
Acquiring positioning data of the pathological area based on a regression algorithm;
generating a recognition result of the pathological image according to the characteristic category and the positioning data;
the step of identifying whether the target pathological feature exists in the feature area diagram through the edge frame network and confirming the area where the target pathological feature exists as a pathological area comprises the following steps:
dividing the characteristic area map into a plurality of first candidate area maps;
identifying whether target pathological features exist in the first candidate region map through an edge frame network;
confirming a first candidate region map with the target pathological features as a candidate pathological region;
dividing the candidate pathological area into a plurality of second candidate area maps;
respectively acquiring Jaccard similarity between corresponding features of a plurality of second candidate region graphs and the target pathological features;
confirming a second candidate region map with the Jaccard similarity higher than a preset similarity threshold as a pathological region;
the step of dividing the candidate pathological area into a plurality of second candidate area maps comprises the following steps:
detecting a cell state within the candidate pathological region;
labeling the foreground and the background of the region of the attached cell in the candidate pathological region, and dividing the labeled attached cell in the region of the attached cell based on a watershed algorithm to obtain a first intermediate processing image;
Performing division-free preprocessing on the complete features in the first intermediate processing image through an edge detection algorithm to obtain a second intermediate processing image;
the second intermediate processed image is divided into a plurality of second candidate region maps based on a region growing algorithm.
2. The pathological image recognition method according to claim 1, wherein the step of subjecting the pathological image to gray processing and subjecting the gray processed pathological image to binarization processing by a maximum inter-class variance method to obtain a binary image comprises:
adjusting the primary colors of the pathological image according to a weighted average method to realize gray scale processing;
detecting the number of pixels corresponding to each gray level in the pathological image after gray level processing to obtain a gray level mapping table;
traversing the gray mapping table based on a maximum inter-class variance method to generate a target gray threshold;
and performing binarization processing on the pathological image subjected to gray level processing according to the target gray level threshold value to obtain a binary image.
3. The pathological image recognition method according to claim 2, wherein the step of generating a target gray threshold value by traversing the gray mapping table based on a maximum inter-class variance method comprises:
Traversing the gray mapping table based on a maximum inter-class variance method to obtain a gray traversing result;
determining an image processing type according to the pathology information corresponding to the pathology image;
confirming corresponding operation of the image processing type through an image processing relation table;
when the image processing type corresponds to the cell edge characteristic strengthening operation, the Matlab platform is combined with the gray level traversing result to adjust the gray level threshold value of the image, so that the cell edge characteristic strengthening operation is completed, and the current gray level threshold value is obtained;
and taking the current gray threshold value as a target gray threshold value.
4. A pathological image recognition method according to any one of claims 1-3, wherein the step of classifying the pathological region corresponding features based on a support vector machine algorithm to obtain feature classes comprises:
detecting the feature dimension level of the corresponding feature of the pathological region;
when the feature dimension level is higher than a preset feature dimension level, performing feature dimension reduction processing on the corresponding features of the pathological region through a principal component analysis method;
classifying the processed corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories.
5. A pathology image recognition method according to any one of claims 1 to 3, wherein said step of denoising said binary image by a neighborhood averaging algorithm, generates a preprocessed pathology image, comprises:
generating an image processing type according to the pathology information corresponding to the pathology image;
determining a neighborhood pixel size and a neighborhood comparison threshold according to the image processing type;
denoising the binary image according to the neighborhood pixel size and the neighborhood comparison threshold value through a neighborhood average algorithm to generate a preprocessing pathological image.
6. A pathology image recognition apparatus, characterized in that the apparatus comprises:
the gray processing module is used for carrying out gray processing on the pathological image, and carrying out binarization processing on the pathological image subjected to gray processing through a maximum inter-class variance method to obtain a binary image;
the image denoising module is used for denoising the binary image through a neighborhood average algorithm to generate a preprocessing pathological image:
the feature extraction module is used for carrying out feature extraction on the preprocessing pathological image based on the single-step multi-frame target detection model to obtain a feature area diagram;
the region confirmation module is used for recognizing whether the target pathological features exist in the feature region diagram through an edge frame network and confirming the region with the target pathological features as a pathological region;
The feature mapping module is used for mapping the pathological region to the feature region map through a region suggestion pooling back propagation algorithm to obtain the corresponding feature of the pathological region;
the feature classification module is used for classifying the corresponding features of the pathological region based on a support vector machine algorithm to obtain feature categories;
the pathology positioning module is used for acquiring positioning data of the pathology area based on a regression algorithm;
the result generation module is used for generating a recognition result of the pathological image according to the characteristic category and the positioning data;
the region confirmation module is further used for dividing the characteristic region map into a plurality of first candidate region maps;
the region confirmation module is further used for identifying whether target pathological features exist in the first candidate region map through an edge frame network;
the region confirmation module is further used for confirming a first candidate region map with the target pathological feature as a candidate pathological region;
the region confirmation module is further used for dividing the candidate pathological region into a plurality of second candidate region graphs;
the region confirmation module is further configured to obtain Jaccard similarities between corresponding features of the plurality of second candidate region graphs and the target pathological feature respectively;
The region confirmation module is further configured to confirm the second candidate region map with the Jaccard similarity higher than a preset similarity threshold as a pathological region;
the region confirmation module is further used for detecting the cell state in the candidate pathological region;
the region confirmation module is further used for marking the foreground and the background of the region where the attached cells in the candidate pathological region belong, and dividing the attached cells in the marked region where the attached cells belong based on a watershed algorithm to obtain a first intermediate processing image;
the region confirmation module is further used for carrying out division-free preprocessing on the complete features in the first intermediate processing image through an edge detection algorithm to obtain a second intermediate processing image;
the region confirmation module is further used for dividing the second intermediate processing image into a plurality of second candidate region graphs based on a region growing algorithm.
7. A pathology image recognition apparatus, characterized in that it comprises: a memory, a processor and a pathology image recognition program stored on the memory and executable on the processor, the pathology image recognition program configured to implement the steps of the pathology image recognition method according to any one of claims 1 to 5.
8. A storage medium having stored thereon a pathology image identification program which, when executed by a processor, implements the steps of the pathology image identification method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310851588.0A CN116580397B (en) | 2023-07-12 | 2023-07-12 | Pathological image recognition method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310851588.0A CN116580397B (en) | 2023-07-12 | 2023-07-12 | Pathological image recognition method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116580397A CN116580397A (en) | 2023-08-11 |
| CN116580397B true CN116580397B (en) | 2023-11-10 |
Family
ID=87545695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310851588.0A Active CN116580397B (en) | 2023-07-12 | 2023-07-12 | Pathological image recognition method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116580397B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111967465A (en) * | 2020-07-07 | 2020-11-20 | 广州金域医学检验中心有限公司 | Method, system, computer device and storage medium for evaluating tumor cell content |
| CN111985536A (en) * | 2020-07-17 | 2020-11-24 | 万达信息股份有限公司 | Gastroscope pathological image classification method based on weak supervised learning |
| WO2021217851A1 (en) * | 2020-04-27 | 2021-11-04 | 平安科技（深圳）有限公司 | Abnormal cell automatic labeling method and apparatus, electronic device, and storage medium |
| CN114972272A (en) * | 2022-06-01 | 2022-08-30 | 东南大学 | Grad-CAM-based segmentation method for new coronary pneumonia lesions |
| CN115294126A (en) * | 2022-10-08 | 2022-11-04 | 南京诺源医疗器械有限公司 | Intelligent cancer cell identification method for pathological image |
| CN115439456A (en) * | 2022-09-18 | 2022-12-06 | 湖南智享未来生物科技有限公司 | Method and device for detecting and identifying object in pathological image |
-
2023
- 2023-07-12 CN CN202310851588.0A patent/CN116580397B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021217851A1 (en) * | 2020-04-27 | 2021-11-04 | 平安科技（深圳）有限公司 | Abnormal cell automatic labeling method and apparatus, electronic device, and storage medium |
| CN111967465A (en) * | 2020-07-07 | 2020-11-20 | 广州金域医学检验中心有限公司 | Method, system, computer device and storage medium for evaluating tumor cell content |
| WO2022007337A1 (en) * | 2020-07-07 | 2022-01-13 | 广州金域医学检验中心有限公司 | Tumor cell content evaluation method and system, and computer device and storage medium |
| CN111985536A (en) * | 2020-07-17 | 2020-11-24 | 万达信息股份有限公司 | Gastroscope pathological image classification method based on weak supervised learning |
| CN114972272A (en) * | 2022-06-01 | 2022-08-30 | 东南大学 | Grad-CAM-based segmentation method for new coronary pneumonia lesions |
| CN115439456A (en) * | 2022-09-18 | 2022-12-06 | 湖南智享未来生物科技有限公司 | Method and device for detecting and identifying object in pathological image |
| CN115294126A (en) * | 2022-10-08 | 2022-11-04 | 南京诺源医疗器械有限公司 | Intelligent cancer cell identification method for pathological image |
Non-Patent Citations (1)
| Title |
|---|
| 淋巴瘤病理图像中R-S细胞的自动提取;陈宏方, 刘秉瀚, 郑智勇;集美大学学报(自然科学版)(02);第142-145页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116580397A (en) | 2023-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109522908B (en) | Image significance detection method based on region label fusion | |
| JP6710135B2 (en) | Cell image automatic analysis method and system | |
| Kothari et al. | Eliminating tissue-fold artifacts in histopathological whole-slide images for improved image-based prediction of cancer grade | |
| Mahapatra | Analyzing training information from random forests for improved image segmentation | |
| Weyn et al. | Automated breast tumor diagnosis and grading based on wavelet chromatin texture description | |
| US8300938B2 (en) | Methods for segmenting objects in images | |
| US20060204953A1 (en) | Method and apparatus for automated analysis of biological specimen | |
| Janssens et al. | CHARISMA: An integrated approach to automatic H&E-stained skeletal muscle cell segmentation using supervised learning and novel robust clump splitting | |
| Yi et al. | Automatic extraction of cell nuclei from H&E-stained histopathological images | |
| Nateghi et al. | Maximized inter-class weighted mean for fast and accurate mitosis cells detection in breast cancer histopathology images | |
| CN112862834A (en) | Image segmentation method based on visual salient region and active contour | |
| WO2013019856A1 (en) | Automated malignancy detection in breast histopathological images | |
| He et al. | Local and global Gaussian mixture models for hematoxylin and eosin stained histology image segmentation | |
| CN111242957A (en) | Data processing method and device, computer storage medium and electronic equipment | |
| Tosta et al. | Computational method for unsupervised segmentation of lymphoma histological images based on fuzzy 3-partition entropy and genetic algorithm | |
| CN111886630A (en) | Three-dimensional cell and tissue image analysis for cellular and subcellular morphological modeling and classification | |
| US20240054639A1 (en) | Quantification of conditions on biomedical images across staining modalities using a multi-task deep learning framework | |
| CN116433704A (en) | Cell nucleus segmentation method based on central point and related equipment | |
| Hu et al. | Breast cancer histopathological images recognition based on two-stage nuclei segmentation strategy | |
| WO2013148485A2 (en) | Detection of tissue regions in microscope slide images | |
| Alilou et al. | Segmentation of cell nuclei in heterogeneous microscopy images: A reshapable templates approach | |
| Agus et al. | Shape analysis of 3D nanoscale reconstructions of brain cell nuclear envelopes by implicit and explicit parametric representations | |
| Sertel et al. | An image analysis approach for detecting malignant cells in digitized H&E-stained histology images of follicular lymphoma | |
| CN111488811B (en) | Face recognition method, device, terminal equipment and computer readable medium | |
| CN116580397B (en) | Pathological image recognition method, device, equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |