CN116309454A - Intelligent pathological image recognition method and device based on lightweight convolution kernel network - Google Patents

Intelligent pathological image recognition method and device based on lightweight convolution kernel network Download PDF

Info

Publication number
CN116309454A
CN116309454A CN202310270552.3A CN202310270552A CN116309454A CN 116309454 A CN116309454 A CN 116309454A CN 202310270552 A CN202310270552 A CN 202310270552A CN 116309454 A CN116309454 A CN 116309454A
Authority
CN
China
Prior art keywords
image sample
image
feature
pathological image
pathological
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.)
Granted
Application number
CN202310270552.3A
Other languages
Chinese (zh)
Other versions
CN116309454B (en
Inventor
邵珠宏
孙智健
尚媛园
赵晓旭
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhaoyang Health Guangzhou Technology Co ltd
Original Assignee
Capital Normal University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Capital Normal University filed Critical Capital Normal University
Priority to CN202310270552.3A priority Critical patent/CN116309454B/en
Publication of CN116309454A publication Critical patent/CN116309454A/en
Application granted granted Critical
Publication of CN116309454B publication Critical patent/CN116309454B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • 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
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
    • G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
    • 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
    • G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
    • 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
    • G06V10/52—Scale-space analysis, e.g. wavelet analysis
    • 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
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
    • G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted 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
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/03—Recognition of patterns in medical or anatomical images
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Radiology & Medical Imaging (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Quality & Reliability (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Image Analysis (AREA)

Abstract

The application provides a pathology image intelligent identification method and device based on a lightweight convolution kernel network, and relates to the field of machine learning, wherein the method comprises the following steps: obtaining a pathological image sample, and carrying out normalization and zero padding operation on the pathological image sample; processing a pathological image sample based on a random Fourier feature transformation approximate Gaussian kernel to obtain nonlinear features of the pathological image sample; encoding nonlinear features through local binarization, dividing the feature map obtained after encoding into non-overlapping image sub-blocks, and calculating a histogram of each image sub-block; and splicing the histograms of the image sub-blocks to generate a histogram set, wherein the histogram set is used as a target feature of the pathological image sample, and the target feature is used for training a support vector machine to generate a support vector machine classifier. According to the model training method and device, training of the model can be effectively completed only by means of small batches of training data, automatic classification of pathological images is achieved, and cost and time are reduced.

Description

Intelligent pathological image recognition method and device based on lightweight convolution kernel network
Technical Field
The application relates to the field of machine learning, in particular to a pathology image intelligent recognition method and device based on a lightweight convolution kernel network.
Background
Pathology images have important significance for diagnosis, treatment and prognosis of diseases, however, computer-aided medical pathology image analysis and diagnosis have been developed due to subjectivity, fatigue and other problems of manual analysis and limitation of high-precision processing capability for large-scale medical data.
By means of computer technology and machine learning algorithm, the pathological image can be automatically processed, characteristic extracted and analyzed, and doctors can be helped to diagnose and treat diseases more objectively and accurately. For example, in cancer pathology image diagnosis, computer-aided diagnosis can rapidly and accurately classify, evaluate prognosis, etc. of cancer, and help to improve diagnosis and treatment level of cancer, and at present, computer-aided medical pathology image diagnosis has become an important research and application direction in pathology field.
However, the recognition of pathological images based on the deep convolutional neural network method often requires a large amount of pathological image data and a complex and trained network structure model to obtain higher recognition accuracy.
Disclosure of Invention
Aiming at the problems, the intelligent pathological image recognition method and device based on the lightweight convolution kernel network are provided, and the technical scheme of the application is as follows:
the first aspect of the application provides a pathology image intelligent identification method based on a lightweight convolution kernel network, which comprises the following steps:
obtaining a pathological image sample, and carrying out normalization and zero padding operation on the pathological image sample;
processing the pathological image sample based on a random Fourier feature transformation approximate Gaussian kernel to obtain nonlinear features of the pathological image sample;
coding the nonlinear characteristics through local binarization, dividing a characteristic diagram obtained after coding into non-overlapping image sub-blocks, and calculating a histogram of each image sub-block;
and splicing the histograms of the image sub-blocks to generate a histogram set, wherein the histogram set is used as a target feature of the pathological image sample, and the target feature is used for training a support vector machine to generate a support vector machine classifier.
Optionally, the obtaining a pathological image sample and performing normalization and zero padding operations on the pathological image sample includes:
selecting a marked pathology image from a published dataset or a hospital provided pathology image as a pathology image sample;
and adjusting the pathological image samples to the same size, and performing four-week zero filling operation to ensure that the obtained first feature matrixes are the same size.
Optionally, a corresponding third feature matrix { T } 31 ,T 32 ,T 33 Largest pooling sizes k 21 ×k 21 、k 22 ×k 22 、k 23 ×k 23 。
Optionally, the encoding the nonlinear feature through local binarization includes:
for an image sub-block of size 3×3, a pixel value g of the center point coordinate is obtained center ；
The pixel value of the surrounding coordinates and the pixel value g of the central point coordinates are compared center Comparing, if the pixel value of the surrounding coordinate is larger than the pixel value of the central point coordinate, setting the pixel value of the coordinate to be 1, otherwise setting the pixel value to be 0;
and arranging the obtained pixel values of the surrounding coordinates into 8-bit binary numbers in a clockwise direction, and converting the 8-bit binary numbers into decimal numbers, wherein the decimal numbers are used as LBP codes of the central point coordinates.
Optionally, the method further comprises:
for a new pathological image, extracting features based on the random Fourier feature transformation approximate Gaussian kernel, and classifying the pathological image by using the support vector machine classifier.
The technical scheme provided by the embodiment of the application at least brings the following beneficial effects:
in the first aspect, a nuclear approximation method and an activation operation are adopted, so that nonlinear characteristics of pathological images are fully excavated;
a second aspect; the method of using the multi-scale filter can improve the receptive field to the pathological image and obtain more effective characteristic representation;
in the third aspect, training of the model can be effectively completed only by small batches of training data, and the problem that a large number of training images are relied on in a depth network model is avoided.
Additional aspects and advantages of the application 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 application.
Drawings
The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:
FIG. 1 is a flow chart illustrating a method for intelligent identification of pathology images based on a lightweight convolutional kernel network, according to an embodiment of the present application;
fig. 2 is a block diagram of a pathology image intelligent recognition device based on a lightweight convolution kernel network according to an embodiment of the present application.
Detailed Description
Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and intended for the purpose of explaining the present application and are not to be construed as limiting the present application.
Fig. 1 is a flowchart of a method for intelligently identifying pathology images based on a lightweight convolution kernel network, according to an embodiment of the present application, including:
step 101, a pathological image sample is obtained, and normalization and zero filling operations are carried out on the pathological image sample.
In the embodiment of the application, the marked pathological image is selected as a pathological image sample so as to facilitate training in the subsequent step, wherein the pathological image can be a public data set or a pathological image provided by a hospital, and the marked content is benign, malignant or other specific categories.
There is no specific limitation on the data set or the hospital.
In order to eliminate the dimension influence among indexes in the pathological image sample, data standardization processing is needed through normalization to solve the comparability among the data indexes, after the original data is subjected to the data standardization processing, each index is in the same order of magnitude and is suitable for comprehensive comparison evaluation, and then, in order to make the size of an output image equal to the size of the input pathological image sample, zero padding is needed for the pathological image sample, and in the application, the specific steps are as follows:
after the pathological image sample is obtained, firstly, carrying out normalization processing on the pathological image sample, adjusting the pathological image sample to be the same size, and carrying out four-week zero filling operation on the pathological image sample so as to ensure that the obtained first feature matrix is the same size.
And 102, processing a pathological image sample based on the approximate Gaussian kernel approximation of the random Fourier feature transformation to obtain nonlinear features of the pathological image sample.
In the embodiment of the application, the random Fourier feature transformation approximately approximates to the Gaussian kernel to be used as a convolution network, so that the nonlinear features of the image can be extracted, and the dimension of the features can be controlled.
First, the pathological image sample X is respectively processed according to three different scales i Traversing pixel by pixel to obtain a set { A } of 3 image sub-blocks 1 ,A 2 ,A 3 Wherein the number of elements in each set is n x m and the three different scales are k 11 ×k 11 、k 12 ×k 12 And k is equal to 13 ×k 13 ；
Secondly, for each image sub-block set, a first feature matrix { T } is obtained by adopting a multi-scale random Fourier feature transformation 11 ,T 12 ,T 13 A first feature matrix having a size of n×m and a dimension of D 1 ；
Then, each first feature matrix is processed through an activation layer to obtain a corresponding second feature matrix { T } 21 ,T 22 ,T 23 }；
Finally, extracting the second feature matrix by a maximum pooling method to obtain a corresponding third feature matrix { T } 31 ,T 32 ,T 33 Largest pooling sizes k 21 ×k 21 、k 22 ×k 22 、k 23 ×k 23 The dimension of the third feature matrix is D 1 。
In the embodiment of the application, through the above process, nonlinear characteristics of the pathological image sample are extracted.
And 103, encoding the nonlinear characteristics through local binarization, dividing the characteristic map obtained after encoding into non-overlapping image sub-blocks, and calculating a histogram of each image sub-block.
In the embodiment of the present application, for an image sub-block with a size of 3×3, a pixel value g of a center point coordinate is obtained center ；
The pixel value g of the surrounding coordinates and the pixel value g of the central point coordinates center Comparing, if the pixel value of the surrounding coordinates is larger than that of the central point coordinates, setting the pixel value of the coordinates to be 1, otherwise setting the pixel value to be 0;
the pixel values of the obtained surrounding coordinates are arranged into 8-bit binary numbers in a clockwise direction, and the 8-bit binary numbers are converted into decimal numbers, wherein the decimal numbers are used as LBP codes of the central point coordinates.
In the embodiment of the application, the coordinates of each coordinate are encoded and converted into integers.
And 104, stitching the histograms of the image sub-blocks to generate a histogram set, wherein the histogram set is used as a target feature of the pathological image sample, and the target feature is used for training a support vector machine to generate a support vector machine classifier.
In the embodiment of the application, after obtaining the histogram of each image sub-block, the histograms of the image sub-blocks are spliced, the spliced histogram set is used as the target feature of the pathological image sample, and the target feature is used for training a support vector machine to generate a support vector machine classifier.
When a new pathological image is input, the characteristics are extracted based on the approximate approximation Gaussian kernel of the random Fourier characteristic transformation, and the pathological image is classified by using a support vector machine classifier.
In one possible embodiment, the new pathology image input is classified and then marked as benign.
In the first aspect of the embodiment of the application, a nuclear approximation method and an activation operation are adopted, so that nonlinear characteristics of pathological images are fully excavated; a second aspect; the method of using the multi-scale filter can improve the receptive field to the pathological image and obtain more effective characteristic representation; in the third aspect, training of the model can be effectively completed only by small batches of training data, and the problem that a large number of training images are relied on in a depth network model is avoided.
Fig. 2 is a block diagram of a pathology image intelligent recognition device 200 based on a lightweight convolution kernel network, which includes a preprocessing module 210, a feature extraction module 220, an encoding module 230, and a training module 240, according to an embodiment of the present application.
The preprocessing module 210 is configured to obtain a pathology image sample, and perform normalization and zero padding operations on the pathology image sample;
the feature extraction module 220 is configured to process the pathological image sample based on the approximate gaussian kernel of the random fourier feature transformation, and obtain nonlinear features of the pathological image sample;
the encoding module 230 is configured to encode the nonlinear feature by local binarization, divide the feature map obtained after encoding into non-overlapping image sub-blocks, and calculate a histogram of each image sub-block;
the training module 240 is configured to stitch the histograms of each image sub-block to generate a histogram set, where the histogram set is used as a target feature of the pathological image sample, and the target feature is used to train a support vector machine to generate a support vector machine classifier.
The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims (6)

1. A pathology image intelligent identification method based on a lightweight convolution kernel network is characterized by comprising the following steps:
obtaining a pathological image sample, and carrying out normalization and zero padding operation on the image sample;
processing the pathological image sample based on a random Fourier feature transformation approximate Gaussian kernel to obtain nonlinear features of the pathological image sample;
coding the nonlinear characteristics through local binarization, dividing a characteristic diagram obtained after coding into non-overlapping image sub-blocks, and calculating a histogram of each image sub-block;
and splicing the histograms of the image sub-blocks to generate a histogram set, wherein the histogram set is used as a target feature of the pathological image sample, and the target feature is used for training a support vector machine to generate a support vector machine classifier.
2. The method of claim 1, wherein the acquiring a pathology image sample and normalizing and zero-filling the pathology image sample comprises:
selecting a marked pathology image from a published dataset or a hospital provided pathology image as a pathology image sample;
and adjusting the pathological image samples to the same size, and performing four-week zero filling operation to ensure that the subsequently obtained first feature matrixes are the same size.
3. The method of claim 2, wherein the processing the pathology image based on the stochastic fourier feature transform approximation gaussian kernel to obtain nonlinear features of the pathology image sample comprises:
for the pathological image sample X according to three different scales i Traversing pixel by pixel to obtain a set { A } of 3 image sub-blocks 1 ,A 2 ,A 3 Wherein the number of elements in each set is n x m and the three different scales are k 11 ×k 11 、k 12 ×k 12 And k is equal to 13 ×k 13 ；
For each image sub-block set, adopting multi-scale random Fourier feature transformation to obtain the first feature matrix { T } 11 ,T 12 ,T 13 The first feature matrix has a size of n×m and a dimension of D 1 ；
Processing each first feature matrix through an activation layer to obtain a corresponding second feature matrix { T } 21 ,T 22 ,T 23 }；
Extracting the second feature matrix by a maximum pooling method to obtain a corresponding third feature matrix { T } 31 ,T 32 ,T 33 Largest pooling sizes k 21 ×k 21 、k 22 ×k 22 、k 23 ×k 23 。
4. A method according to claim 3, wherein said encoding said non-linear features by local binarization comprises:
for an image sub-block of size 3×3, a pixel value g of the center point coordinate is obtained center ；
The pixel value of the surrounding coordinates and the pixel value g of the central point coordinates are compared center Comparing, if the pixel value of the surrounding coordinate is larger than the pixel value of the central point coordinate, setting the pixel value of the coordinate to be 1, otherwise setting the pixel value to be 0;
and arranging the obtained pixel values of the surrounding coordinates into 8-bit binary numbers in a clockwise direction, and converting the 8-bit binary numbers into decimal numbers, wherein the decimal numbers are used as LBP codes of the central point coordinates.
5. The method according to claim 1, wherein the method further comprises:
for a new pathological image, extracting features based on the random Fourier feature transformation approximate Gaussian kernel, and classifying the pathological image by using the support vector machine classifier.
6. Pathology image intelligent identification device based on lightweight convolution kernel network, its characterized in that includes:
the preprocessing module is used for acquiring a pathological image sample and carrying out normalization and zero padding operation on the pathological image sample;
the feature extraction module is used for processing the pathological image sample based on the approximate Gaussian kernel of the random Fourier feature transformation to obtain the nonlinear feature of the pathological image sample;
the coding module is used for coding the nonlinear characteristics through local binarization, dividing a characteristic diagram obtained after coding into non-overlapping image sub-blocks, and calculating a histogram of each image sub-block;
and the training module is used for splicing the histograms of the image sub-blocks to generate a histogram set, wherein the histogram set is used as a target feature of the pathological image sample, and the target feature is used for training a support vector machine to generate a support vector machine classifier.
CN202310270552.3A 2023-03-16 2023-03-16 Intelligent pathological image recognition method and device based on lightweight convolution kernel network Active CN116309454B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310270552.3A CN116309454B (en) 2023-03-16 2023-03-16 Intelligent pathological image recognition method and device based on lightweight convolution kernel network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310270552.3A CN116309454B (en) 2023-03-16 2023-03-16 Intelligent pathological image recognition method and device based on lightweight convolution kernel network

Publications (2)

Publication Number Publication Date
CN116309454A true CN116309454A (en) 2023-06-23
CN116309454B CN116309454B (en) 2023-09-19

Family

ID=86801025

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310270552.3A Active CN116309454B (en) 2023-03-16 2023-03-16 Intelligent pathological image recognition method and device based on lightweight convolution kernel network

Country Status (1)

Country Link
CN (1) CN116309454B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2817553C1 (en) * 2023-10-13 2024-04-16 Общество с ограниченной ответственностью "Малое инновационное предприятие "ЭКОМ" (ООО "МИП "ЭКОМ") Method of imaging impregnation of capillary-porous objects

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019223147A1 (en) * 2018-05-23 2019-11-28 平安科技（深圳）有限公司 Liver canceration locating method and apparatus, and storage medium
WO2020077866A1 (en) * 2018-10-17 2020-04-23 平安科技（深圳）有限公司 Moire-based image recognition method and apparatus, and device and storage medium
CN115546862A (en) * 2022-09-14 2022-12-30 江苏科技大学 Expression recognition method and system based on cross-scale local difference depth subspace characteristics

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019223147A1 (en) * 2018-05-23 2019-11-28 平安科技（深圳）有限公司 Liver canceration locating method and apparatus, and storage medium
WO2020077866A1 (en) * 2018-10-17 2020-04-23 平安科技（深圳）有限公司 Moire-based image recognition method and apparatus, and device and storage medium
CN115546862A (en) * 2022-09-14 2022-12-30 江苏科技大学 Expression recognition method and system based on cross-scale local difference depth subspace characteristics

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
FELIX JUEFEI-XU ETC.: ""Subspace-Based Discrete Transform Encoded Local Binary Patterns Representations for Robust Periocular Matching on NIST’s Face Recognition Grand Challenge"", 《IEEE TRANSACTIONS ON IMAGE PROCESSING》, vol. 23, no. 8, XP011552863, DOI: 10.1109/TIP.2014.2329460 *
胡敏等: ""基于非对称局部梯度编码的人脸表情识别"", 《中国图象图形学报》, vol. 20, no. 10 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2817553C1 (en) * 2023-10-13 2024-04-16 Общество с ограниченной ответственностью "Малое инновационное предприятие "ЭКОМ" (ООО "МИП "ЭКОМ") Method of imaging impregnation of capillary-porous objects

Also Published As

Publication number Publication date
CN116309454B (en) 2023-09-19

Similar Documents

Publication Publication Date Title
CN111798464A (en) Lymphoma pathological image intelligent identification method based on deep learning
CN111524144B (en) Intelligent lung nodule diagnosis method based on GAN and Unet network
CN110930378B (en) Emphysema image processing method and system based on low data demand
CN113610859B (en) Automatic thyroid nodule segmentation method based on ultrasonic image
CN115359066B (en) Focus detection method and device for endoscope, electronic device and storage medium
CN116758336A (en) Medical image intelligent analysis system based on artificial intelligence
Huynh et al. Plant identification using new architecture convolutional neural networks combine with replacing the red of color channel image by vein morphology leaf
CN111931817A (en) Pellet ore phase identification method and device
CN110136113B (en) Vagina pathology image classification method based on convolutional neural network
CN111401434B (en) Image classification method based on unsupervised feature learning
CN116309454B (en) Intelligent pathological image recognition method and device based on lightweight convolution kernel network
Thapa et al. Deep learning for breast cancer classification: Enhanced tangent function
CN116958126A (en) Method and system for detecting injection molding of mold based on computer vision
Salama et al. Enhancing Medical Image Quality using Neutrosophic Fuzzy Domain and Multi-Level Enhancement Transforms: A Comparative Study for Leukemia Detection and Classification
CN113177602B (en) Image classification method, device, electronic equipment and storage medium
CN113870194B (en) Breast tumor ultrasonic image processing device with fusion of deep layer characteristics and shallow layer LBP characteristics
Guo et al. Thyroid nodule ultrasonic imaging segmentation based on a deep learning model and data augmentation
Kalaiselvi et al. Improved classification of brain tumor in MR images using RNN classification framework
CN112906690A (en) License plate segmentation model training method, license plate segmentation method and related device
CN117173485B (en) Intelligent classification system method and system for lung cancer tissue pathological images
CN109993201B (en) Image processing method, device and readable storage medium
CN111414956B (en) Multi-example learning identification method for fuzzy mode in lung CT image
CN114612718B (en) Small sample image classification method based on graph structural feature fusion
CN113936165B (en) CT image processing method, terminal and computer storage medium
Subha et al. Analysis of Deep Learning based Optimization Techniques for Oral Cancer Detection

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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20240402

Address after: 509 Kangrui Times Square, Keyuan Business Building, 39 Huarong Road, Gaofeng Community, Dalang Street, Longhua District, Shenzhen City, Guangdong Province, 518110

Patentee after: Shenzhen lizhuan Technology Transfer Center Co.,Ltd.

Country or region after: China

Address before: 100037 No. 105 West Third Ring Road North, Beijing, Haidian District

Patentee before: Capital Normal University

Country or region before: China

TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20240416

Address after: Room 501, Room 401, No. 50 Zhongshan 7th Road, Jinhua Street, Liwan District, Guangzhou City, Guangdong Province, 510000 (for office use only)

Patentee after: Zhaoyang Health (Guangzhou) Technology Co.,Ltd.

Country or region after: China

Address before: 509 Kangrui Times Square, Keyuan Business Building, 39 Huarong Road, Gaofeng Community, Dalang Street, Longhua District, Shenzhen City, Guangdong Province, 518110

Patentee before: Shenzhen lizhuan Technology Transfer Center Co.,Ltd.

Country or region before: China