CN113392895A - Knee joint cartilage damage detection method and system - Google Patents

Knee joint cartilage damage detection method and system Download PDF

Info

Publication number
CN113392895A
CN113392895A CN202110645160.1A CN202110645160A CN113392895A CN 113392895 A CN113392895 A CN 113392895A CN 202110645160 A CN202110645160 A CN 202110645160A CN 113392895 A CN113392895 A CN 113392895A
Authority
CN
China
Prior art keywords
cartilage
cartilage damage
region
knee joint
image information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110645160.1A
Other languages
Chinese (zh)
Inventor
怀晓晨
穆红章
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.)
Lingyu Yinnuo Beijing Technology Co ltd
Original Assignee
Lingyu Yinnuo Beijing Technology Co ltd
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 Lingyu Yinnuo Beijing Technology Co ltd filed Critical Lingyu Yinnuo Beijing Technology Co ltd
Priority to CN202110645160.1A priority Critical patent/CN113392895A/en
Publication of CN113392895A publication Critical patent/CN113392895A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • 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
    • 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
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10088—Magnetic resonance imaging [MRI]
    • 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
    • G06T2207/30008—Bone

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biophysics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Biomedical Technology (AREA)
  • Evolutionary Biology (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

The embodiment of the application discloses a knee joint cartilage injury detection method and system, wherein the method comprises the following steps: acquiring knee joint MRI image information, and preprocessing the image information; extracting cartilage damage areas layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume area; extracting cartilage damage interested areas layer by layer to obtain cartilage damage three-dimensional areas; extracting three-dimensional interesting region characteristics, training a classifier network by adopting a neural network method, and determining an optimal model; and inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection, and determining the cartilage damage degree. The method has good robustness, is rapid and convenient, and is suitable for assisting diagnosis of doctors.

Description

Knee joint cartilage damage detection method and system
Technical Field
The embodiment of the application relates to the technical field of medical detection, in particular to a knee joint cartilage injury detection method and system.
Background
Osteoarthritis (OA) is a common chronic, progressive, highly disabling degenerative disease of the joints. The diagnosis of the cartilage damage of the knee joint provides important reference significance for clinic. MRI is a multi-parameter imaging examination method, and matrix component changes of knee joint cartilage can be found in time before the knee joint cartilage is subjected to pathological morphological changes, so that early diagnosis is carried out on the cartilage damage degree of the knee joint. With the continuous development and research of computer software and the use of various magnetic resonance sequences with high signal-to-noise ratio and high resolution, more and more image algorithms are applied to the automatic detection of cartilage of the knee joint, and a non-invasive and radiation-free automatic detection method for cartilage damage of the knee joint based on MRI is gradually formed. Has important clinical value in the disease course monitoring and curative effect evaluation of OA.
The knee joint cartilage is complicated in grading, the severity of the knee joint cartilage cannot be accurately evaluated, and particularly the knee joint cartilage is difficult to identify in an early disease stage, so that a DenseNet detection cartilage injury method is provided to provide reference decisions for doctors in an auxiliary diagnosis process. And judging the diseased condition of the cartilage by using a DenseNet method through the marked cartilage area, thereby realizing the automatic cartilage damage detection of the knee joint.
However, the traditional algorithm needs to adjust different threshold values to adapt to variable scenes, the extraction effect is difficult to guarantee, and a large amount of manual setting is needed.
Disclosure of Invention
Therefore, the embodiment of the application provides a knee joint cartilage injury detection method and system, which are good in robustness, fast and convenient and suitable for assisting diagnosis of doctors.
In order to achieve the above object, the embodiments of the present application provide the following technical solutions:
according to a first aspect of embodiments of the present application, there is provided a knee joint cartilage damage detection method, the method including:
acquiring knee joint MRI image information, and preprocessing the image information;
extracting cartilage damage areas layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume area;
extracting cartilage damage interested areas layer by layer to obtain cartilage damage three-dimensional areas;
extracting three-dimensional interesting region characteristics, training a classifier network by adopting a neural network method, and determining an optimal model;
and inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection, and determining the cartilage damage degree.
Optionally, the extracting a cartilage damage region of interest layer by layer to obtain a cartilage damage three-dimensional region includes:
marking a region of interest ROI; rescaling all ROI (region of interest) by using linear interpolation, marking the middle size as a uniform size, and carrying out category marking on the diseased cartilage region, wherein the category marking comprises health, semi-injury and complete injury; and dividing the training set, the verification set and the test set according to a set proportion.
Optionally, the extracting the three-dimensional region of interest features, training a classifier network by using a neural network method, and determining an optimal model includes:
training an interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final marked interested area; training a classifier network by adopting a neural network method, establishing a connection relation between different layers by adopting a DenseNet structure, and generating a candidate model; the different layers include a Bottleneck layer, a transformation layer and a growth rate; and determining the optimal model through a plurality of times of iterative training based on the loss convergence condition.
Optionally, the step of calculating the raw loss function of the region of interest network and the classifier network comprises:
the normalized probability p of softmax is calculated as followsi：
Figure BDA0003108963720000031
xi＝xi-max(x1,…，xn)
Wherein i is the ith sample, j is the jth sample, and n is the nth sample; x is the number ofiIs the probability of the ith sample;
loss was calculated according to the following formula:
Loss＝-wk*log pk
wherein p iskIs the probability of sample k;
and dynamically optimizing the weight combination according to the image quality and the applicable scene so as to ensure that the Loss function obtains the minimum value.
Optionally, the inputting the MRI image information of the knee joint into the optimal model, performing cartilage damage detection, and determining a cartilage damage degree includes:
inputting the knee joint MRI image information into the optimal model, preprocessing the image information, and extracting a region with cartilage;
inputting the image data with cartilage into a regression model, and outputting a marked region of interest;
inputting the extracted region-of-interest images into a classifier network, outputting the probability of each category, selecting the category with the highest probability, and outputting a diagnosis result; if the probability of outputting 0 is the maximum, the cartilage is healthy; if the probability of outputting 1 is maximum, cartilage semi-damage is carried out; if the probability of outputting 2 is the maximum, cartilage is completely damaged.
According to a second aspect of embodiments of the present application, there is provided a knee joint cartilage damage detection system, the system comprising:
the image information acquisition module is used for acquiring knee joint MRI image information and preprocessing the image information;
the cartilage three-dimensional volume region extraction module is used for extracting cartilage damage regions layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume region;
the cartilage damage three-dimensional region extraction module is used for extracting cartilage damage regions of interest layer by layer to obtain a cartilage damage three-dimensional region;
the model training module is used for extracting the characteristics of the three-dimensional region of interest, training a classifier network by adopting a neural network method and determining an optimal model;
and the cartilage damage detection module is used for inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection and determining the cartilage damage degree.
Optionally, the cartilage damage three-dimensional region extraction module is specifically configured to:
marking a region of interest ROI;
rescaling all ROI (region of interest) by using linear interpolation, marking the middle size as a uniform size, and carrying out category marking on the diseased cartilage region, wherein the category marking comprises health, semi-injury and complete injury;
and dividing the training set, the verification set and the test set according to a set proportion.
Optionally, the model training module is specifically configured to:
training an interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final marked interested area;
training a classifier network by adopting a neural network method, establishing a connection relation between different layers by adopting a DenseNet structure, and generating a candidate model; the different layers include a Bottleneck layer, a transformation layer and a growth rate;
and determining the optimal model through a plurality of times of iterative training based on the loss convergence condition.
According to a third aspect of embodiments herein, there is provided an apparatus comprising: the device comprises a data acquisition device, a processor and a memory; the data acquisition device is used for acquiring data; the memory is to store one or more program instructions; the processor is configured to execute one or more program instructions to perform the method of any of the first aspect.
According to a fourth aspect of embodiments herein, there is provided a computer-readable storage medium having one or more program instructions embodied therein for performing the method of any of the first aspects.
In summary, the embodiment of the present application provides a knee joint cartilage damage detection method and system, by acquiring knee joint MRI image information and preprocessing the image information; extracting cartilage damage areas layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume area; extracting cartilage damage interested areas layer by layer to obtain cartilage damage three-dimensional areas; extracting three-dimensional interesting region characteristics, training a classifier network by adopting a neural network method, and determining an optimal model; and inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection, and determining the cartilage damage degree. The method has good robustness, is rapid and convenient, and is suitable for assisting diagnosis of doctors.
Drawings
In order to more clearly illustrate the implementation of the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the implementation or the prior art will be briefly described below. It should be apparent that the drawings in the following description are merely exemplary, and that other embodiments can be derived from the drawings provided by those of ordinary skill in the art without inventive effort.
The structures, ratios, sizes, and the like shown in the present specification are only used for matching with the contents disclosed in the present specification, so that those skilled in the art can understand and read the present disclosure, and do not limit the conditions that the embodiments of the present application can be implemented, so that the present disclosure has no technical significance, and any structural modifications, changes in the ratio relationship, or adjustments of the size should still fall within the scope that the technical contents disclosed in the embodiments of the present application can cover without affecting the efficacy and the achievable purpose that the embodiments of the present application can be implemented.
Fig. 1 is a schematic flow chart of a knee joint cartilage damage detection method provided in the embodiment of the present application;
FIG. 2 is a block diagram of a neural network method training classifier network provided in an embodiment of the present application;
FIGS. 3a and 3b are schematic diagrams of the accuracy and loss convergence provided by the embodiments of the present application;
fig. 4 is a block diagram of a knee joint cartilage damage detection system according to an embodiment of the present application.
Detailed Description
Other advantages and features of the embodiments of the present application will become apparent to those skilled in the art from the following description, wherein it is to be understood that the embodiments of the present application are described in connection with the particular illustrative embodiments thereof, and that the embodiments of the present application are not limited to the particular embodiments disclosed herein. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without any creative effort belong to the protection scope of the embodiments in the present application.
Fig. 1 shows a flow of a knee joint cartilage damage detection method provided in an embodiment of the present application, which specifically includes the following steps:
step 101: acquiring knee joint MRI image information, and preprocessing the image information;
step 102: extracting cartilage damage areas layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume area;
step 103: extracting cartilage damage interested areas layer by layer to obtain cartilage damage three-dimensional areas;
step 104: extracting three-dimensional interesting region characteristics, training a classifier network by adopting a neural network method, and determining an optimal model;
step 105: and inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection, and determining the cartilage damage degree.
In a possible embodiment, in step 103, the extracting a cartilage damage region of interest layer by layer to obtain a cartilage damage three-dimensional region includes:
marking a region of interest ROI; rescaling all ROI (region of interest) by using linear interpolation, marking the middle size as a uniform size, and carrying out category marking on the diseased cartilage region, wherein the category marking comprises health, semi-injury and complete injury; and dividing the training set, the verification set and the test set according to a set proportion.
In a possible implementation manner, in step 104, the extracting three-dimensional region of interest features, training a classifier network by using a neural network method, and determining an optimal model includes:
training an interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final marked interested area; training a classifier network by adopting a neural network method, establishing a connection relation between different layers by adopting a DenseNet structure, and generating a candidate model; the different layers include a Bottleneck layer, a transformation layer and a growth rate; and determining the optimal model through a plurality of times of iterative training based on the loss convergence condition.
In one possible embodiment, the step of calculating the raw loss function of the region of interest network and the classifier network comprises:
the normalized probability p of softmax is calculated according to the following equations (1) and (2)i：
Figure BDA0003108963720000061
xi＝xi-max(x1,…，xn) (2)
Wherein i is the ith sample, j is the jth sample, and n is the nth sample; x is the number ofiIs the probability of the ith sample;
the Loss is calculated according to the following formula (3):
Loss＝-wk*log pk (3)
wherein p iskIs the probability of sample k;
and dynamically optimizing the weight combination according to the image quality and the applicable scene so as to ensure that the Loss function obtains the minimum value.
In a possible implementation manner, in step 105, the inputting the MRI image information of the knee joint into the optimal model to perform cartilage damage detection and determine the degree of cartilage damage includes:
inputting the knee joint MRI image information into the optimal model, preprocessing the image information, and extracting a region with cartilage; inputting the image data with cartilage into a regression model, and outputting a marked region of interest; inputting the extracted region-of-interest images into a classifier network, outputting the probability of each category, selecting the category with the highest probability, and outputting a diagnosis result; if the probability of outputting 0 is the maximum, the cartilage is healthy; if the probability of outputting 1 is maximum, cartilage semi-damage is carried out; if the probability of outputting 2 is the maximum, cartilage is completely damaged.
Aiming at the problems that knee joint cartilage is complicated in grading, the severity of the knee joint cartilage cannot be accurately evaluated, and particularly the knee joint cartilage is difficult to identify in the early stage of illness, the DenseNet detection method for cartilage injury is provided, can provide guiding significance for clinical diagnosis of knee joint cartilage injury, and provides theoretical basis for disease course monitoring and curative effect evaluation. When the knee joint MRI image is subjected to cartilage detection, a cascade model is adopted, so that the fine cartilage results existing in a low-contrast and micro-target mode in the full-image visual field can be effectively identified and extracted, and meanwhile, the loss function of the cascade model is optimized, so that the model has higher robustness.
The following provides a further detailed explanation of the embodiments of the knee joint cartilage damage detection method provided in the embodiments of the present application. The method specifically comprises the following steps:
s1: and collecting knee joint MRI image information, and preprocessing image data.
Wherein, step S1 specifically includes:
s11: preprocessing of the knee joint MR sequence original graph: during the processing of the data, the image data of the missing section is washed away.
S12: knee joint image data normalization: and (4) carrying out normalization processing on the data, and uniformly processing the data into 512 × 512 sizes.
In the process of processing the data, cleaning the image data lacking slices, acquiring complete image data, then carrying out normalization processing on the data, uniformly processing each layer of data into 512 multiplied by 512, and providing the normalization processing of the pixel gray level of the DICOM data.
S2: and extracting cartilage damage areas layer by layer to generate a cartilage three-dimensional volume area under an MR scanning T2 sequence.
The effective volume of the knee joint cartilage MR T2 image is observed to be different from 290 × 300 × 21 to 320 × 320 × 60, so that the knee joint cartilage region in the image is matched by a method cv2.matchtemplate () to make statistical atlas template matching, and the middle part region 320 × 320 × 32 of the image is cut according to the matching result.
S3: and extracting cartilage injury interested areas layer by layer. Generating a three-dimensional region of cartilage damage;
wherein step S3 specifically includes:
s31: labeling the regions of interest, the extracted ROI size varied from 54 × 46 × 2 to 124 × 136 × 6, so all ROI regions of interest were rescaled using linear interpolation, labeling uniformly the middle size as 90 × 90 (slice height × slice width) size, and class labeling the diseased cartilage regions, labeling the data as class 3: healthy, semi-injured, completely injured.
S32: data were divided into training set, validation set and test set according to 8:1:1 ratio.
S4: and extracting the characteristics of the three-dimensional region of interest, training a classifier and generating a model.
Wherein, step S4 specifically includes:
s41: extracting the characteristics of the region of interest;
and training the interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final example graph for marking the interested area.
S42: training a classifier;
the neural network method is adopted to train the classifier network (as shown in figure 2), the DenseNet structure is adopted, the connection relation between different layers is established, the characteristics of the different layers are fully utilized, the gradient disappearance problem is further reduced, the network discomfort problem is deepened, and the training effect is good. In addition, the use of the Bottleneck layer, the transformation layer and the smaller growth rate enables the network to be narrowed, the parameters to be reduced, the overfitting is effectively inhibited, and meanwhile, the calculation amount is reduced. And finally generating a candidate model.
S43, generating a model;
and observing the loss convergence condition through multiple iterative training according to the trained figures 3a and 3b, and finally selecting the optimal model with smaller loss.
Wherein, in S41 and S42, the step of calculating the original loss function of the regression model and the classification model comprises:
a. calculating the normalized probability of softmax according to the formulas (1) and (2);
b. calculating the loss, then: loss ═ logpkAnd k is sample label.
Due to the serious imbalance between the cartilage pixels and the background pixels, the softmax Loss function is optimized in the embodiment, and when the Loss is calculated, the labels of different classes are multiplied by different weights w, so that the formula (3) is given: loss ═ wk*log pk(ii) a In the formula, pkIs the probability that a sample belongs to k; according to the image quality and the applicable scene, the weight combination is dynamically optimized, so that the Loss function obtains the minimum value, the problem that the model cannot be converged to a better position due to imbalance of the foreground and the background is solved, and the optimal effect of the regression model and the classification model is ensured.
S5: and selecting an optimal model, detecting the original data, and judging the cartilage damage degree.
Wherein step S5 specifically includes:
s51: selecting an optimal model, inputting original image data, preprocessing the original image, extracting a region with cartilage,
s52: putting the image data with cartilage into a regression model, outputting the marked interesting region,
s53: inputting the extracted region-of-interest images into a classifier, outputting the probability of each category, and selecting the category with the highest probability, such as the category with the highest probability of outputting 0, which represents that the cartilage is healthy; the probability of outputting 1 is the maximum, which represents the cartilage semi-damage; the probability of outputting 2 is the maximum, which represents that the cartilage is completely damaged, and the result is used as a diagnosis result.
In particular practice, 90 cases were tested by the method of the examples of the present application, finally achieving an accuracy of 91.17%, with missing partial loss cases due to less obvious cartilage structure. Therefore, the embodiment of the application is effective in detecting most knee joint cartilage, can realize automatic detection and greatly improves the efficiency.
In summary, the embodiment of the present application provides a knee joint cartilage damage detection method, which includes acquiring knee joint MRI image information and preprocessing the image information; extracting cartilage damage areas layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume area; extracting cartilage damage interested areas layer by layer to obtain cartilage damage three-dimensional areas; extracting three-dimensional interesting region characteristics, training a classifier network by adopting a neural network method, and determining an optimal model; and inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection, and determining the cartilage damage degree. The method has good robustness, is rapid and convenient, and is suitable for assisting diagnosis of doctors.
Based on the same technical concept, the embodiment of the present application further provides a block diagram of a knee joint cartilage injury detection system, as shown in fig. 4, the system includes:
the image information acquisition module 401 is configured to acquire knee joint MRI image information and preprocess the image information;
a cartilage three-dimensional volume region extraction module 402, configured to perform cartilage damage region layer-by-layer extraction on the preprocessed image information to obtain a knee joint cartilage three-dimensional volume region;
a cartilage damage three-dimensional region extraction module 403, configured to extract cartilage damage regions of interest layer by layer to obtain a cartilage damage three-dimensional region;
a model training module 404, configured to extract features of a three-dimensional region of interest, train a classifier network by using a neural network method, and determine an optimal model;
and a cartilage damage detection module 405, configured to input the knee joint MRI image information into the optimal model, perform cartilage damage detection, and determine a cartilage damage degree.
In a possible implementation manner, the cartilage damage three-dimensional region extraction module 403 is specifically configured to:
marking a region of interest ROI; rescaling all ROI (region of interest) by using linear interpolation, marking the middle size as a uniform size, and carrying out category marking on the diseased cartilage region, wherein the category marking comprises health, semi-injury and complete injury; and dividing the training set, the verification set and the test set according to a set proportion.
In a possible implementation manner, the model training module 404 is specifically configured to:
training an interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final marked interested area; training a classifier network by adopting a neural network method, establishing a connection relation between different layers by adopting a DenseNet structure, and generating a candidate model; the different layers include a Bottleneck layer, a transformation layer and a growth rate; and determining the optimal model through a plurality of times of iterative training based on the loss convergence condition.
Based on the same technical concept, an embodiment of the present application further provides an apparatus, including: the device comprises a data acquisition device, a processor and a memory; the data acquisition device is used for acquiring data; the memory is to store one or more program instructions; the processor is configured to execute one or more program instructions to perform the method.
Based on the same technical concept, the embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium contains one or more program instructions, and the one or more program instructions are used for executing the method.
In the present specification, each embodiment of the method is described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. Reference is made to the description of the method embodiments.
It should be noted that although the operations of the methods of the embodiments of the present application are depicted in the drawings in a particular order, this does not require or imply that these operations must be performed in this particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.
Although the present application provides method steps as in embodiments or flowcharts, additional or fewer steps may be included based on conventional or non-inventive approaches. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. When an apparatus or client product in practice executes, it may execute sequentially or in parallel (e.g., in a parallel processor or multithreaded processing environment, or even in a distributed data processing environment) according to the embodiments or methods shown in the figures. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus 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 apparatus. Without further limitation, the presence of additional identical or equivalent elements in a process, method, article, or apparatus that comprises the recited elements is not excluded.
The units, devices, modules, etc. set forth in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. Of course, in implementing the present application, the functions of each module may be implemented in one or more software and/or hardware, or a module implementing the same function may be implemented by a combination of a plurality of sub-modules or sub-units, and the like. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may therefore be considered as a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, classes, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
From the above description of the embodiments, it is clear to those skilled in the art that the present application can be implemented by software plus necessary general hardware platform. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, or the like, and includes several instructions for enabling a computer device (which may be a personal computer, a mobile terminal, a server, or a network device) to execute the method according to the embodiments or some parts of the embodiments of the present application.
The embodiments in the present specification are described in a progressive manner, and the same or similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. The application is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
The above-mentioned embodiments are further described in detail for the purpose of illustrating the invention, and it should be understood that the above-mentioned embodiments are only illustrative of the present invention and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (10)

1. A method for detecting cartilage damage in a knee joint, the method comprising:
acquiring knee joint MRI image information, and preprocessing the image information;
extracting cartilage damage areas layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume area;
extracting cartilage damage interested areas layer by layer to obtain cartilage damage three-dimensional areas;
extracting three-dimensional interesting region characteristics, training a classifier network by adopting a neural network method, and determining an optimal model;
and inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection, and determining the cartilage damage degree.
2. The method of claim 1, wherein extracting the cartilage damage region of interest layer by layer to obtain a cartilage damage three-dimensional region comprises:
marking a region of interest ROI;
rescaling all ROI (region of interest) by using linear interpolation, marking the middle size as a uniform size, and carrying out category marking on the diseased cartilage region, wherein the category marking comprises health, semi-injury and complete injury;
and dividing the training set, the verification set and the test set according to a set proportion.
3. The method of claim 1, wherein the extracting three-dimensional region of interest features, training a classifier network using a neural network method, and determining an optimal model comprises:
training an interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final marked interested area;
training a classifier network by adopting a neural network method, establishing a connection relation between different layers by adopting a DenseNet structure, and generating a candidate model; the different layers include a Bottlenecklayer, a transformation layer, and a growth rate;
and determining the optimal model through a plurality of times of iterative training based on the loss convergence condition.
4. The method of claim 3, wherein the step of computing the raw loss functions of the region of interest network and the classifier network comprises:
the normalized probability p of softmax is calculated as followsi：
Figure FDA0003108963710000021
xi＝xi-max(x1,…，xn)
Wherein i is the ith sample, j is the jth sample, and n is the nth sample; x is the number ofiIs the probability of the ith sample;
loss was calculated according to the following formula:
Loss＝-wk*log pk
wherein p iskIs the probability of sample k;
and dynamically optimizing the weight combination according to the image quality and the applicable scene so as to ensure that the Loss function obtains the minimum value.
5. The method of claim 1, wherein the inputting the knee joint MRI image information into the optimal model for cartilage damage detection and determining the degree of cartilage damage comprises:
inputting the knee joint MRI image information into the optimal model, preprocessing the image information, and extracting a region with cartilage;
inputting the image data with cartilage into a regression model, and outputting a marked region of interest;
inputting the extracted region-of-interest images into a classifier network, outputting the probability of each category, selecting the category with the highest probability, and outputting a diagnosis result; if the probability of outputting 0 is the maximum, the cartilage is healthy; if the probability of outputting 1 is maximum, cartilage semi-damage is carried out; if the probability of outputting 2 is the maximum, cartilage is completely damaged.
6. A knee joint cartilage damage detection system, the system comprising:
the image information acquisition module is used for acquiring knee joint MRI image information and preprocessing the image information;
the cartilage three-dimensional volume region extraction module is used for extracting cartilage damage regions layer by layer according to the preprocessed image information to obtain a knee joint cartilage three-dimensional volume region;
the cartilage damage three-dimensional region extraction module is used for extracting cartilage damage regions of interest layer by layer to obtain a cartilage damage three-dimensional region;
the model training module is used for extracting the characteristics of the three-dimensional region of interest, training a classifier network by adopting a neural network method and determining an optimal model;
and the cartilage damage detection module is used for inputting the knee joint MRI image information into the optimal model, performing cartilage damage detection and determining the cartilage damage degree.
7. The system of claim 6, wherein the cartilage damage three-dimensional region extraction module is specifically configured to:
marking a region of interest ROI;
rescaling all ROI (region of interest) by using linear interpolation, marking the middle size as a uniform size, and carrying out category marking on the diseased cartilage region, wherein the category marking comprises health, semi-injury and complete injury;
and dividing the training set, the verification set and the test set according to a set proportion.
8. The system of claim 6, wherein the model training module is specifically configured to:
training an interested area network by adopting a deep learning neural network, and performing regression simulation on the size of the interested area to obtain a final marked interested area;
training a classifier network by adopting a neural network method, establishing a connection relation between different layers by adopting a DenseNet structure, and generating a candidate model; the different layers include a Bottleneck layer, a transformation layer and a growth rate;
and determining the optimal model through a plurality of times of iterative training based on the loss convergence condition.
9. An apparatus, characterized in that the apparatus comprises: the device comprises a data acquisition device, a processor and a memory;
the data acquisition device is used for acquiring data; the memory is to store one or more program instructions; the processor, configured to execute one or more program instructions to perform the method of any of claims 1-5.
10. A computer-readable storage medium having one or more program instructions embodied therein for performing the method of any of claims 1-5.
CN202110645160.1A 2021-06-09 2021-06-09 Knee joint cartilage damage detection method and system Pending CN113392895A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110645160.1A CN113392895A (en) 2021-06-09 2021-06-09 Knee joint cartilage damage detection method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110645160.1A CN113392895A (en) 2021-06-09 2021-06-09 Knee joint cartilage damage detection method and system

Publications (1)

Publication Number Publication Date
CN113392895A true CN113392895A (en) 2021-09-14

Family

ID=77620048

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110645160.1A Pending CN113392895A (en) 2021-06-09 2021-06-09 Knee joint cartilage damage detection method and system

Country Status (1)

Country Link
CN (1) CN113392895A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023241031A1 (en) * 2022-06-15 2023-12-21 北京长木谷医疗科技有限公司 Deep learning-based three-dimensional intelligent diagnosis method and system for osteoarthritis

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023241031A1 (en) * 2022-06-15 2023-12-21 北京长木谷医疗科技有限公司 Deep learning-based three-dimensional intelligent diagnosis method and system for osteoarthritis

Similar Documents

Publication Publication Date Title
Zhao et al. A deep learning model integrating FCNNs and CRFs for brain tumor segmentation
Abdullah et al. Automatic detection and classification of knee osteoarthritis using deep learning approach
Villalobos-Castaldi et al. A fast, efficient and automated method to extract vessels from fundus images
Wu et al. Texture feature based automated seeded region growing in abdominal MRI segmentation
Myint et al. Analysis on leg bone fracture detection and classification using X-ray images
CN109949280B (en) Image processing method, image processing apparatus, device storage medium, and growth evaluation system
US11227387B2 (en) Multi-stage brain tumor image processing method and system
Ribeiro et al. Handling inter-annotator agreement for automated skin lesion segmentation
Dandu et al. Brain and pancreatic tumor segmentation using SRM and BPNN classification
He et al. Mammographic image segmentation and risk classification based on mammographic parenchymal patterns and geometric moments
Masmoudi et al. LBPV descriptors-based automatic ACR/BIRADS classification approach
Xu et al. Classifying mammographic breast density by residual learning
Oyeranmi et al. Detection of fracture bones in X-ray images categorization
Kanchanamala et al. Optimization-enabled hybrid deep learning for brain tumor detection and classification from MRI
Jerbi et al. Automatic classification of ultrasound thyroids images using vision transformers and generative adversarial networks
Sindhuja et al. A Novel Technique for Ensembled Learning based on Convolution Neural Network
Behara et al. Artificial intelligence in medical diagnostics: A review from a South African context
CN113724185B (en) Model processing method, device and storage medium for image classification
CN113392895A (en) Knee joint cartilage damage detection method and system
Manjula Gururaj Rao et al. Data Analytics Based Multimodal System for Fracture Identification and Verification in CBIR Domain
Amiriebrahimabadi et al. A Comprehensive Survey of Multi-Level Thresholding Segmentation Methods for Image Processing
Ma et al. Computer aided mass detection in mammography with temporal change analysis
Kaur et al. A survey on medical image segmentation
Chhabra et al. Comparison of different edge detection techniques to improve quality of medical images
Setty et al. Survey on Features and Techniques used for Bone Fracture Detection and Classification

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