CN113436145A - Bone age determination method and device based on artificial intelligence and electronic equipment - Google Patents
Bone age determination method and device based on artificial intelligence and electronic equipment Download PDFInfo
- Publication number
- CN113436145A CN113436145A CN202110566880.9A CN202110566880A CN113436145A CN 113436145 A CN113436145 A CN 113436145A CN 202110566880 A CN202110566880 A CN 202110566880A CN 113436145 A CN113436145 A CN 113436145A
- Authority
- CN
- China
- Prior art keywords
- region
- interest
- neural network
- image
- joint
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- 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
- G06T2207/10116—X-ray image
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The application provides a bone age determination method and device based on artificial intelligence, electronic equipment and a computer readable storage medium; the method comprises the following steps: determining a region of interest in a digitized X-ray (DR) image, the DR image comprising a skeletal image of a limb; acquiring a grade evaluation value of the region of interest; and determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest. By the method and the device, the bone age can be determined quickly and accurately.
Description
Technical Field
The present application relates to artificial intelligence technology, and in particular, to a method and an apparatus for determining bone age based on artificial intelligence, an electronic device, and a computer-readable storage medium.
Background
Bone age (also called bone age), which is more reflective of a person's actual growth and development than actual age, may be used to characterize different stages of bone development. Bone age assessment has a key role in diagnosis of endocrine diseases, prediction of adult height, assessment of treatment effects, and the like.
In the related art, the method for evaluating bone age has at least the problems of poor evaluation accuracy and long evaluation time.
Disclosure of Invention
The embodiment of the application provides a bone age determination method and device based on artificial intelligence, an electronic device and a computer-readable storage medium, which can improve the accuracy of bone age assessment and reduce the time required by bone age assessment.
The technical scheme of the embodiment of the application is realized as follows:
in a first aspect, an embodiment of the present application provides a bone age determination method based on artificial intelligence, including:
determining a region of interest in a Digital Radiography (DR) image, the DR image comprising a skeletal image of a limb; acquiring a grade evaluation value of the region of interest; and determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest.
In the above solution, the determining the region of interest in the DR image includes:
and inputting the DR image into a first neural network model, and detecting the metacarpophalangeal joints in the skeleton image by using the first neural network model to obtain a first region of interest corresponding to the metacarpophalangeal joints.
In the above solution, before the determining the region of interest in the DR image, the method further includes:
generating a first set of training samples for training the first neural network model;
obtaining metacarpophalangeal joint characteristics of each sample in the first training sample set;
and training the first neural network model by taking the metacarpophalangeal joint characteristics as granularity, so that the first neural network model can predict a first region of interest corresponding to the metacarpophalangeal joint.
In the above solution, the determining the region of interest in the DR image includes:
and inputting the DR image into a second neural network model, and segmenting the wrist joint in the skeleton image by using the second neural network model to obtain a second region of interest corresponding to the wrist joint.
In the above scheme, the segmenting the wrist joint in the bone image by using the second neural network model to obtain a second region of interest corresponding to the wrist joint includes:
segmenting a wrist joint in the bone image by using the second neural network model, and determining a contour line comprising the wrist joint;
and determining the area in the first shape minimum enclosing frame corresponding to the contour line as the second region of interest.
In the above solution, before the determining the region of interest in the DR image based on the bone image, the method further comprises:
generating a second set of training samples for training the second neural network model;
acquiring wrist joint features of each sample in the second training sample set;
training the second neural network model with the wrist joint features as granularity, so that the second neural network model can predict a second region of interest corresponding to a wrist joint.
In the foregoing solution, the acquiring a rank evaluation value of the region of interest includes:
inputting the region of interest into a third neural network model, and acquiring a level evaluation value of a metacarpophalangeal joint and/or a wrist joint corresponding to the region of interest by using the third neural network model; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
In the foregoing solution, before the obtaining the rank evaluation value of the region of interest, the method further includes:
generating a third set of training samples for training the third neural network model;
obtaining an interested area corresponding to a metacarpophalangeal joint and/or a wrist joint in each sample in the third training sample set;
and training a third neural network model by taking the characteristics of the interested areas corresponding to the metacarpophalangeal joints and/or the wrist joints as granularity, so that the second neural network model can predict the grade evaluation values corresponding to the metacarpophalangeal joints and/or the wrist joints.
In the above scheme, the determining a bone age corresponding to the bone image based on the level evaluation value of the region of interest includes:
calculating the sum of the grade evaluation values of all the interested areas;
and determining the bone age corresponding to the bone image based on the sum of the grade evaluation values.
In a second aspect, an embodiment of the present application provides a bone age determination model training method, where the method includes:
generating a sample set of historical DR images, the historical DR images comprising a skeletal image of a limb;
acquiring joint features of each sample in the sample set;
training an interested region neural network model by taking the joint features as granularity, so that the interested region neural network model can predict an interested region corresponding to the joint;
obtaining an interested area corresponding to a joint in each sample;
training a grade evaluation model by taking the characteristics of the interested area of the joint as granularity, so that the grade evaluation model can predict a grade evaluation value corresponding to the joint; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
In the above scheme, the training of the neural network model of the region of interest with the joint features as the granularity includes:
training a first neural network model by taking the metacarpophalangeal joint features in the joint features as granularity, so that the first neural network model can predict a first region of interest corresponding to the metacarpophalangeal joint;
and/or training a second neural network model by taking the wrist joint features in the joint features as granularity, so that the second neural network model can predict a second region of interest corresponding to the wrist joint.
In a third aspect, an embodiment of the present application provides an artificial intelligence-based bone age determination device, including:
a region-of-interest determination module for determining a region of interest in a DR image, the DR image comprising a skeleton image of a limb;
a grade evaluation value acquisition module, configured to acquire a grade evaluation value of the region of interest;
and the bone age determining module is used for determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest.
In a fourth aspect, an embodiment of the present application provides a bone age determination model training device, where the device includes:
a generation module for generating a sample set of historical DR images, the historical DR images including a skeletal image of a limb;
the joint feature acquisition module is used for acquiring the joint features of each sample in the sample set;
the first training module is used for training an interested region neural network model by taking the joint features as granularity, so that the interested region neural network model can predict an interested region corresponding to a joint;
a region of interest acquisition module for acquiring a region of interest corresponding to a joint in each sample
The second training module is used for training a grade evaluation model by taking the characteristics of the interested area of the joint as granularity, so that the grade evaluation model can predict a grade evaluation value corresponding to the joint; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
In a fifth aspect, an embodiment of the present application provides an electronic device, including:
a memory for storing executable instructions;
and the processor is used for realizing the bone age determining method based on artificial intelligence provided by the embodiment of the application or realizing the bone age determining model training method provided by the embodiment of the application when the executable instructions stored in the memory are executed.
In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores executable instructions for implementing, when executed by a processor, the method for determining a bone age based on artificial intelligence provided in the embodiment of the present application, or implementing the method for training a model for determining a bone age provided in the embodiment of the present application.
The bone age determining method based on artificial intelligence provided by the embodiment of the application determines an interested area in a DR image, wherein the DR image comprises a bone image of limbs; acquiring a grade evaluation value of the region of interest; and determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest. Because the process of the bone age determining method based on artificial intelligence is completed by the electronic equipment, compared with the prior art that joints are manually searched according to DR images, the joints are manually analyzed according to clinical knowledge, and bones are manually classified and graded according to the characteristics of the bones so as to determine the bone age, the bone age evaluating time is greatly shortened. In addition, the bone age is determined based on artificial intelligence, namely the bone age of a new DR image is determined based on a large number of historical DR images, so that the bone age can be determined more accurately by the embodiment of the application.
Drawings
FIG. 1 is a schematic diagram of an architecture of an artificial intelligence-based bone age determination system provided by an embodiment of the present application;
FIG. 2 is a schematic structural diagram of a server applying an artificial intelligence-based bone age determination method according to an embodiment of the present application;
FIG. 3 is a schematic view of an alternative process flow of a method for determining bone age based on artificial intelligence according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a region of interest input to a third neural network model provided by an embodiment of the present application;
FIG. 5 is a schematic diagram of an alternative process flow for training a third neural network model according to an embodiment of the present application;
FIG. 6 is a schematic overall process flow diagram of a method for determining bone age based on artificial intelligence according to an embodiment of the present application;
FIG. 7 is a schematic view of an alternative process flow of a training method for a bone age determination model according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of a training device for a bone age determination model according to an embodiment of the present application.
Detailed Description
In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.
In the following description, references to the terms "first", "second", and the like, are only to distinguish similar objects and do not denote a particular order, but rather the terms "first", "second", and the like may be used interchangeably with the order specified, where permissible, to enable embodiments of the present application described herein to be practiced otherwise than as specifically illustrated or described herein.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
Before further detailed description of the embodiments of the present application, terms and expressions referred to in the embodiments of the present application will be described, and the terms and expressions referred to in the embodiments of the present application will be used for the following explanation.
1) Artificial Intelligence (AI) is a theory, method and technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human Intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results.
2) Region Of Interest (ROI): in machine vision and image processing, a region to be processed is outlined in a square, circle, ellipse, or the like from a processed image is called a region of interest. The region of interest is obtained by various operators (operators) and functions commonly used in machine vision software such as Halcon, OpenCV, Matlab and the like, and the image is processed in the next step.
3) The DR system, i.e. a direct digital radiography system, is composed of an electronic cassette, a scan controller, a system controller, an image monitor, etc., and directly converts X-ray photons into a digital image through the electronic cassette, which is a direct digital radiography in a broad sense. In the narrow sense, direct digital radiography (ddr) generally refers to digital radiography using a direct image conversion technique of a flat panel detector, and is a real direct digital X-ray radiography system. Mainly divided into an amorphous silicon flat plate DR, an amorphous selenium flat plate DR and a CCD DR according to the type of the detector; the frame structure is divided into a suspension DR and a column (UC arm) DR.
The following technical solutions exist for bone age assessment in the related art:
1. G-P (Greulich and Pyle) spectroscopy; G-P spectroscopy compares a patient's DR images with recent standard X-rays. However, the accuracy of this method can only be half a year, and the subjective nature of the physician can lead to significant differences in the assessment results.
2. TW3(Tanner-Whitehouse 3) bone age assessment method; the TW3 bone age assessment method was based on a scoring system that allowed bone age estimation accuracy within one month. Specifically, the reviewer would first identify 20 bones (e.g., 13 metacarpophalangeal bones and 7 carpal bones) each having a classification stage, replace each classification stage with a score, calculate an overall score and convert it to bone age.
Although the TW3 bone age assessment method is more accurate than the G-P atlas method, the TW3 bone age assessment method requires a long time, a complicated process flow, and requires at least 30 minutes for a reviewer to manually assess. Even with computer-aided detection systems, the rating of each bone is still dependent on the interpretation of the reviewer, which inevitably leads to differences in bone age assessment results from different reviewers.
Aiming at the problems of long time required by bone age assessment, poor accuracy and the like in the methods provided by the related technology, the embodiment of the application provides a bone age determination method based on artificial intelligence, a device, electronic equipment and a computer readable storage medium, which can solve the problems of long time required by bone age assessment and poor accuracy.
An exemplary application of the electronic device provided in the embodiment of the present application is described below, and the electronic device provided in the embodiment of the present application may be implemented as a server. In the following, an exemplary application will be explained when the electronic device is implemented as a server.
The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud computing services. The terminal may be, but is not limited to, a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and the like. The terminal and the server may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present application is not limited herein.
The so-called artificial intelligence cloud Service is also generally called AIaaS (AI as a Service, chinese). The method is a service mode of an artificial intelligence platform, and particularly, the AIaaS platform splits several types of common AI services and provides independent or packaged services at a cloud. This service model is similar to the one opened in an AI theme mall: all developers can access one or more artificial intelligence services provided by the platform by means of an API (application programming interface), and part of the qualified developers can also use an AI framework and an AI infrastructure provided by the platform to deploy and operate and maintain the self-dedicated cloud artificial intelligence services
Referring to fig. 1, fig. 1 is a schematic diagram of an architecture of an artificial intelligence-based bone age determining system provided in an embodiment of the present application, which may be used to determine bone ages corresponding to bone images in a DR image. In the bone age determination system in which the terminal 400 is connected to the server 200 through the network 300, the network 300 may be a wide area network or a local area network, or a combination of both, in response to receiving a bone age determination request from the terminal 400, the functions of the bone age determination system may be implemented based on the respective modules in the server 200, in the process that a user uses a client, a terminal 400 reports collected DR images and bone ages corresponding to bone images in the DR images to a database 600 as training sample data, the training sample data are data from different individuals reported by each terminal, a server 200 trains a neural network model based on the obtained training data, and in response to the server 200 receiving a bone age determination request of the terminal 400, an interested region determination module 2551 in the server 200 determines an interested region in the DR images according to the trained neural network model; the grade evaluation value acquisition module 2552 acquires the grade evaluation value of the region of interest according to the trained neural network model; and the bone age determination module 2553 determines the bone age corresponding to the bone image based on the correspondence between the sum of the level evaluation values of the region of interest and the bone age.
As an example, the terminal 400 may install and run a related application. The application refers to a program corresponding to the server and providing local service for the client. Here, the local service may include, but is not limited to: displaying the DR image and sending the DR image to a server. The terminal in the embodiment of the present application may include, but is not limited to, any electronic product based on an intelligent operating system, which can perform human-computer interaction with a user through an input device such as a keyboard, a virtual keyboard, a touch pad, a touch screen, and a voice control device, such as a smart phone, a tablet computer, a personal computer, and the like. Smart operating systems include, but are not limited to, any operating system that enriches device functionality by providing various mobile applications to a mobile device, such as: android, IOS, and the like.
It is understood that the architecture of the artificial intelligence based bone age determination system of fig. 1 is only a partial exemplary implementation in the embodiments of the present application, and the architecture of the artificial intelligence based bone age determination system in the embodiments of the present application includes, but is not limited to, the architecture of the artificial intelligence based bone age determination system shown in fig. 1.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a server 200 applying an artificial intelligence-based bone age determination method according to an embodiment of the present application, where the server 200 shown in fig. 2 includes: at least one processor 210, memory 250, and at least one network interface 220. The various components in server 200 are coupled together by a bus system 240. It is understood that the bus system 240 is used to enable communications among the components. The bus system 240 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 240 in fig. 2.
The Processor 210 may be an integrated circuit chip having Signal processing capabilities, such as a general purpose Processor, a Digital Signal Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like, wherein the general purpose Processor may be a microprocessor or any conventional Processor, or the like.
The memory 250 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard disk drives, optical disk drives, and the like. Memory 250 optionally includes one or more storage devices physically located remotely from processor 210.
In some embodiments, memory 250 is capable of storing data, examples of which include programs, modules, and data structures, or a subset or superset thereof, to support various operations, as exemplified below.
An operating system 251 including system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks;
a network communication module 252 for communicating to other computing devices via one or more (wired or wireless) network interfaces 220, exemplary network interfaces 220 including: bluetooth, wireless compatibility authentication (WiFi), and Universal Serial Bus (USB), among others.
In some embodiments, the artificial intelligence based information recommendation device provided by the embodiments of the present application may be implemented in software, and fig. 2 shows an artificial intelligence based bone age determination device 255 stored in a memory 250, which includes a plurality of modules of a bone age determination system, where the modules may be software in the form of programs and plug-ins, and the like, and include the following software modules: a region-of-interest determining module 2551, a rank evaluation value acquiring module 2552, and a bone age determining module 2553, which are logical and thus can be arbitrarily combined or further divided according to the functions implemented, and the functions of the respective modules will be described below.
The method for determining bone age based on artificial intelligence provided by the embodiments of the present application will be described in connection with exemplary applications and implementations of a bone age determining system provided by the embodiments of the present application, which includes a training phase and an application phase.
First, an application of the model in the artificial intelligence-based bone age determination method provided in the embodiment of the present application will be described. Referring to fig. 3, fig. 3 is a schematic view of an alternative processing flow of the artificial intelligence-based bone age determining method according to the embodiment of the present application, which will be described with reference to steps S101 to S103 shown in fig. 3.
Step S101, determining a region of interest in a DR image, wherein the DR image comprises a skeleton image of a limb.
In some embodiments, first, an artificial intelligence based bone age determination device acquires a DR image. The DR image may include a skeletal image of a limb; the limb may be a hand or foot, etc. Taking the limb as a hand as an example, the skeleton image included in the DR image may be an image in which the fingers are open upward, or an image in which the fingers are open downward. Then, an artificial intelligence based bone age determination apparatus determines a region of interest in the DR image. The region of interest may be a region that needs to be processed and is delineated from the DR image in a square, circle, ellipse, or the like.
In some embodiments, taking as an example that the DR image includes a skeletal image of a hand, the regions of interest may include a first region of interest corresponding to a metacarpophalangeal joint and a second region of interest corresponding to a wrist joint.
And aiming at the first region of interest, inputting the DR image into a first neural network model by using an artificial intelligence-based bone age determining device, and detecting the metacarpophalangeal joints of the hand by using the first neural network model to obtain the first region of interest corresponding to the metacarpophalangeal joints. As an example, when detecting a metacarpophalangeal joint of a hand, 3 metacarpals, 8 phalanges, 1 ulna and 1 radius may be detected, and a range of the detected metacarpophalangeal joint is outlined in a rectangle or other shape, and an area within the outlined range is the first region of interest. In specific implementation, the number of the metacarpophalangeal bones is not limited to 13 in the above example, or may be other numbers, and the embodiment of the present application is not limited.
In some embodiments, the first neural network model may also be referred to as a target detection model, which may be, for example, a fast-RCNN target detection model. The artificial intelligence based bone age determination apparatus may also train the first neural network model prior to determining the region of interest in the DR image. As an example, an artificial intelligence based bone age determination apparatus generates a first set of training samples for training the first neural network model; the first training sample set may be dicom images of the hand, which are normalized by the maximum and/or minimum values to obtain png format images. Labeling all sample images in the first training sample set, such as labeling a surrounding frame (namely an interested area) of a metacarpophalangeal joint; the metacarpophalangeal joints in the sample image can also be identified, and if the first metacarpophalangeal joint is identified as 1, the second metacarpophalangeal joint is identified as 2, and the like, the first metacarpophalangeal joint and the second metacarpophalangeal joint can be divided based on the positions of the fingers where the metacarpophalangeal joints are located and the metacarpophalangeal joints in the fingers. As an example, the size of the anchor box (anchor size) in the fast-RCNN object detection model may be set according to actual conditions, and as an example, the size of the anchor box may be set to 8 × 8 pixels, or 16 × 16 pixels, or 32 × 32 pixels, or 64 × 64 pixels, etc.,
thus, the first neural network model is trained through the known images of the metacarpophalangeal joints and the interested areas corresponding to the metacarpophalangeal joints, so that the first neural network model can determine the first interested areas corresponding to the metacarpophalangeal joints according to the images of the metacarpophalangeal joints in the new skeleton images; as an example, the first neural network model can determine the type, confidence, and the like of the metacarpophalangeal bones from the images of the metacarpophalangeal joints in the new bone image. The confidence coefficient is used for representing the credibility of the first region of interest obtained based on the first neural network model, and the higher the confidence coefficient is, the higher the credibility of the first region of interest is indicated. To improve the accuracy of the first neural network model, bone images of different age groups may be included in the first set of training samples.
And aiming at a second region of interest, inputting the DR image into a second neural network model by using an artificial intelligence-based bone age determining device, and segmenting the wrist joint in the bone image by using the second neural network model to obtain a second region of interest corresponding to the wrist joint. As an example, when detecting a wrist joint of a hand, 7 pieces of carpal bones may be segmented; if the background, foreground or mask of 7 carpal bones except for the pisiform bones is used as a true value (groudtruth) of segmentation, and the wrist DR image of the hand is used as the input of the second neural network model, the second interested area corresponding to the wrist joint can be output. When a second region of interest is determined, firstly, a contour line of the wrist joint is drawn, and then a region in the first minimum enclosing frame in the shape corresponding to the contour line is determined as the second region of interest; as an example, the first shape may be a rectangle. In specific implementation, the number of wrist joints is not limited to the 7 pieces illustrated above, or may be another number, and the embodiment of the present application is not limited.
In some embodiments, the second neural network model may also be referred to as a semantic segmentation model, which may be, for example, a deplaybv 3 semantic segmentation model. The artificial intelligence based bone age determination apparatus may also train a second neural network model prior to determining the region of interest in the DR image. As an example, an artificial intelligence based bone age determination apparatus generates a second set of training samples for training the second neural network model; the second training sample set may be dicom images of the hand, which are normalized by the maximum and/or minimum values to obtain png format images. Labeling all sample images in the second training sample set, such as labeling a bounding box (i.e., a region of interest) of the wrist joint; wrist joints in the sample image may also be identified, and if a first type of wrist joint is identified as a, a second type of wrist joint is identified as B, etc., the first type of wrist joint and the second type of wrist joint may be divided based on the position of the wrist joint. In this way, the second neural network model is trained through the known wrist joint image and the wrist joint corresponding region of interest, so that the second neural network model can determine a second region of interest corresponding to the wrist joint according to the wrist joint image in the new bone image; the second region of interest may also become a mask for the wrist joint; as an example, the second neural network model can be based on an image of the wrist joint in the new bone image, and can also determine the category, confidence, and the like of the wrist. And the confidence coefficient is used for representing the credibility of the second region of interest obtained based on the second neural network model, and the higher the confidence coefficient is, the higher the credibility of the second region of interest is indicated. To improve the accuracy of the second neural network model, bone images of different age groups may be included in the second set of training samples.
In some embodiments, the first set of training samples and the second set of training samples may be the same set of training samples or may be different sets of training samples. The same training sample set may mean that the samples in the first training sample set and the second training sample set are identical; different sets of training samples may mean that the samples in the first set of training samples and the second set of training samples are not identical or are not identical.
Step S102, obtaining a grade evaluation value of the region of interest.
In some embodiments, the artificial intelligence-based bone age determining device inputs the region of interest into a third neural network model, and a level evaluation value of a metacarpophalangeal joint and/or a wrist joint corresponding to the region of interest is obtained by using the third neural network model; and the grade evaluation value is used for representing the bone age corresponding to the bone image. Wherein, each bone in each age stage or each classification stage corresponds to a grade evaluation value, such as 0.1 for 0-1 year old, 0.2 for 2-3 years old, etc.
As an example, the level evaluation value corresponding to the bone may be determined according to the TW3 rating standard.
In some embodiments, the schematic diagram of the region of interest input to the third neural network model may include a first region of interest corresponding to a metacarpophalangeal joint and a second region of interest corresponding to a wrist joint, as shown in fig. 4.
In some embodiments, the artificial intelligence based bone age determination device may pre-train the third neural network model. In particular, an optional process flow for training the third neural network model, as shown in fig. 5, includes:
step S102a, generating a third training sample set for training the third neural network model.
In some embodiments, the third set of training samples is the same as at least one of the first set of training samples and the second set of training samples, and the third set of training samples may also be different from both the first set of training samples and the second set of training samples.
Step S102b, obtaining a region of interest corresponding to the metacarpophalangeal joint and/or the wrist joint in each sample in the third training sample set.
In some embodiments, the artificial intelligence based bone age determining apparatus may acquire the region of interest corresponding to the metacarpophalangeal joint and/or the wrist joint in each sample of the third training sample set based on the same manner as in step S101.
Step S102c, training a third neural network model by using the features of the region of interest corresponding to the metacarpophalangeal joint and/or the wrist joint as granularity, so that the third neural network model can predict a level evaluation value corresponding to the metacarpophalangeal joint and/or the wrist joint.
In some embodiments, the third neural network model may be a classification model, by way of example, the third neural network model may be a Resnet classification model; the artificial intelligence-based bone age determination device may identify a region of interest in the third training sample set, and train the third neural network model using a rank evaluation value of the region of interest as a label. In this way, the third neural network model is trained through the known region of interest corresponding to the metacarpophalangeal joint and the known region of interest corresponding to the wrist joint, so that the third neural network model can determine the grade evaluation value corresponding to the metacarpophalangeal joint and/or the wrist joint according to the region of interest in the new bone image. In order to improve the accuracy of the third neural network model, bone images of different age groups and regions of interest corresponding to metacarpophalangeal joints and wrist joints of different classes may be included in the third training sample set.
Step S103, determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest.
In some embodiments, the artificial intelligence-based bone age determination means calculates a sum of rating values of all regions of interest, and determines the bone age corresponding to the bone image based on the sum of rating values. As an example, the sum of the level evaluation values has a correspondence with the bone age; according to the calculated grade grades and the sum, the corresponding bone age can be determined.
As an example, the sum of the grade evaluation values of all the first regions of interest may be calculated, and the bone age corresponding to the corresponding metacarpophalangeal joints may be determined according to the correspondence between the sum of the grade evaluation values and the bone age; and calculating the sum of grade evaluation values of all the second interested areas, and determining the bone age corresponding to the corresponding wrist joint according to the corresponding relation between the sum of the grade evaluation values and the bone age. Then, determining the bone age corresponding to the bone image based on the bone age corresponding to the metacarpophalangeal bone joint and the bone age corresponding to the wrist joint; as an example, an average value of the bone age corresponding to the metacarpophalangeal joint and the bone age corresponding to the wrist joint may be calculated, and the average value may be used as the bone age corresponding to the bone image; as an example, the larger of the bone age corresponding to the metacarpophalangeal joint and the bone age corresponding to the wrist joint may be used as the bone age corresponding to the bone image; as an example, the smaller of the bone age corresponding to the metacarpophalangeal joint and the bone age corresponding to the wrist joint may be used as the bone age corresponding to the bone image.
In specific implementation, the correspondence between the sum of the ranking evaluation values and the bone age can be determined by a TW3 bone age score table.
An overall processing flow diagram of the artificial intelligence based bone age determination method provided in the embodiment of the present application can be shown in fig. 6, and includes: 1) a DR image including an image of a bone is acquired. 2) A first region of interest corresponding to a metacarpophalangeal joint is obtained by using a target Detection model (Detection module), and a second region of interest corresponding to a wrist joint is obtained by using a semantic Segmentation model (Segmentation module). 3) Respectively calculating the Level evaluation values (Level & score) of the first region of interest and the second region of interest by using a Classification model (Classification module); for example, the metacarpophalangeal joint and the wrist joint may be classified by using a classification model, and the grade evaluation value under the classification may be determined according to the TW3 age scoring rule used in clinical diagnosis. 4) Calculating the sum of grade evaluation values of all the first interested areas, and determining the bone age (bone-R) corresponding to the metacarpophalangeal joints according to the corresponding relation between the sum of the grade evaluation values and the bone age; and calculating the sum of grade evaluation values of all the second interested areas, and determining the bone age (bone-C) corresponding to the corresponding wrist joint according to the corresponding relation between the sum of the grade evaluation values and the bone age. 5) Determining the bone age corresponding to the bone image based on the bone age corresponding to the metacarpophalangeal bone joint and the bone age corresponding to the wrist joint; as an example, an average value of the bone age corresponding to the metacarpophalangeal joint and the bone age corresponding to the wrist joint may be calculated, and the average value may be used as the bone age corresponding to the bone image; as an example, the larger of the bone age corresponding to the metacarpophalangeal joint and the bone age corresponding to the wrist joint may be used as the bone age corresponding to the bone image; as an example, the smaller of the bone age corresponding to the metacarpophalangeal joint and the bone age corresponding to the wrist joint may be used as the bone age corresponding to the bone image.
The bone age determining method based on artificial intelligence provided by the embodiment of the application is characterized in that a neural network model for determining the bone age is trained in advance, wherein the neural network model can comprise a first neural network model, a second neural network model and a grade evaluation model (a third neural network model); the bone age corresponding to the bone image included in the DR image can be obtained by using the DR image as the input of the neural network model. The process of determining the bone age in the embodiment of the application is completed by electronic equipment, and compared with the related technology that joints are manually searched according to DR images, the joints are manually analyzed according to clinical knowledge, and the bones are manually classified and graded according to the characteristics of the bones so as to determine the bone age, the bone age assessment time is greatly shortened. Since the neural network model for determining the bone age is obtained based on a large number of historical DR images, the embodiments of the present application can determine the bone age more accurately.
An embodiment of the present application further provides a bone age determination model training method, and an optional processing flow of the bone age determination model training method, as shown in fig. 7, may include:
step S301, generating a historical DR image sample set, wherein the historical DR image comprises a skeleton image of a limb.
In some embodiments, the historical DR image sample set may be at least one of the first training sample set and the second training sample set described in step S101.
Step S302, acquiring the joint features of each sample in the sample set.
In some embodiments, the joint features may include metacarpophalangeal joint features and wrist joint features.
Step S303, training a region-of-interest neural network model by taking the joint features as granularity, so that the region-of-interest neural network model can predict a region of interest corresponding to the joint.
In some embodiments, the neural network model may include the first neural network model and the second neural network model and the third neural network model in the above embodiments. The first neural network model and the second neural network model may be independent neural network models trained based on different historical DR images, or may be neural network models having an association relationship. As an example, the association relationship may be that the first training sample set corresponding to the first neural network model is the same as the training sample set corresponding to the second neural network model.
In some embodiments, a first neural network model is trained with a metacarpophalangeal joint feature of the joint features as granularity, so that the first neural network model can predict a first region of interest corresponding to a metacarpophalangeal joint; and training a second neural network model by taking the wrist joint features in the joint features as granularity, so that the second neural network model can predict a second region of interest corresponding to the wrist joint.
In some embodiments, the process of training the neural network model may be the same as the process of training the first neural network model and the second neural network model in the above embodiments, and details are not repeated here.
And step S304, acquiring a region of interest corresponding to the joint in each sample.
In some embodiments, a region of interest corresponding to a joint in each sample may be obtained based on the outputs of the first and second neural network models.
Step S305, training a grade evaluation model by taking the characteristics of the interested area of the joint as granularity, so that the grade evaluation model can predict a grade evaluation value corresponding to the joint; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
In some embodiments, the process of training the rating model may be the same as the process of training the third neural network model in the above embodiments, and is not described here again.
In the embodiment of the application, the model for determining the bone age is obtained based on a large amount of historical DR images, so that the bone age corresponding to the bone image included in any DR image can be quickly and accurately obtained by using the trained model.
Continuing with the exemplary structure of the artificial intelligence based bone age determination device 255 provided by the embodiments of the present application as software modules, in some embodiments, as shown in fig. 2, the software modules stored in the artificial intelligence based bone age determination device 255 of the memory 250 may include: a region-of-interest determination module 2551 for determining a region of interest in a DR image, the DR image comprising a skeletal image of a limb; a grade evaluation value obtaining module 2552, configured to obtain a grade evaluation value of the region of interest; and a bone age determining module 2553, configured to determine a bone age corresponding to the bone image based on the level evaluation value of the region of interest.
In some embodiments, the region of interest determining module 2551 is configured to input the DR image to a first neural network model, and detect a metacarpophalangeal joint in the bone image by using the first neural network model to obtain a first region of interest corresponding to the metacarpophalangeal joint.
In some embodiments, the region of interest determination module 2551 is further configured to generate a first set of training samples for training the first neural network model;
obtaining metacarpophalangeal joint characteristics of each sample in the first training sample set;
and training the first neural network model by taking the metacarpophalangeal joint characteristics as granularity, so that the first neural network model can predict a first region of interest corresponding to the metacarpophalangeal joint.
In some embodiments, the region of interest determining module 2551 is configured to input the DR image into a second neural network model, and segment a wrist joint in the bone image by using the second neural network model to obtain a second region of interest corresponding to the wrist joint.
In some embodiments, the region of interest determining module 2551 is configured to segment a wrist joint in the bone image using the second neural network model, determine a contour line including the wrist joint; and determining the area in the first shape minimum enclosing frame corresponding to the contour line as the second region of interest.
In some embodiments, the region of interest determination module 2551 is further configured to generate a second set of training samples for training the second neural network model;
acquiring wrist joint features of each sample in the second training sample set;
training the second neural network model with the wrist joint features as granularity, so that the second neural network model can predict a second region of interest corresponding to a wrist joint.
In some embodiments, the level evaluation value obtaining module 2552 is configured to input the region of interest into a third neural network model, and obtain a level evaluation value of a metacarpophalangeal joint and/or a wrist joint corresponding to the region of interest by using the third neural network model; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
In some embodiments, the rank evaluation value obtaining module 2552 is further configured to generate a third training sample set for training the third neural network model;
obtaining an interested area corresponding to a metacarpophalangeal joint and/or a wrist joint in each sample in the third training sample set;
and training a third neural network model by taking the characteristics of the interested areas corresponding to the metacarpophalangeal joints and/or the wrist joints as granularity, so that the second neural network model can predict the grade evaluation values corresponding to the metacarpophalangeal joints and/or the wrist joints.
In some embodiments, the bone age determination module 2553 is configured to calculate a sum of all the rating values of the region of interest; and determining the bone age corresponding to the bone image based on the sum of the grade evaluation values.
Continuing with the exemplary architecture of the bone age determination model training apparatus 400 implemented as software modules provided by the embodiments of the present application, in some embodiments, as shown in fig. 8, the bone age determination model training apparatus 400 may include:
a generating module 401, configured to generate a historical DR image sample set, where the historical DR image includes a skeleton image of a limb;
a joint feature obtaining module 402, configured to obtain a joint feature of each sample in the sample set;
a first training module 403, configured to train a region-of-interest neural network model with the joint features as granularity, so that the region-of-interest neural network model can predict a region of interest corresponding to a joint;
a region-of-interest obtaining module 404 for obtaining a region of interest corresponding to a joint in each sample
A second training module 405, configured to train a level evaluation model with the feature of the region of interest of the joint as a granularity, so that the level evaluation model can predict a level evaluation value corresponding to the joint; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
It should be noted that the description of the apparatus in the embodiment of the present application is similar to the description of the method embodiment, and has similar beneficial effects to the method embodiment, and therefore, the description is not repeated. The inexhaustible technical details of the artificial intelligence-based bone age determination device and the bone age determination model training device provided by the embodiment of the application can be understood from the description of any one of the drawings in fig. 1 to 7.
Embodiments of the present application provide a computer-readable storage medium storing executable instructions, which when executed by a processor, will cause the processor to perform the artificial intelligence based bone age determination method provided by embodiments of the present application, or will cause the processor to perform the bone age determination model training method provided by embodiments of the present application.
In some embodiments, the computer-readable storage medium may be memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories.
In some embodiments, executable instructions may be written in any form of programming language (including compiled or interpreted languages), in the form of programs, software modules, scripts or code, and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.
By way of example, executable instructions may be deployed to be executed on one computing device or on multiple computing devices at one site or distributed across multiple sites and interconnected by a communication network.
The above description is only an example of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, and improvement made within the spirit and scope of the present application are included in the protection scope of the present application.
Claims (15)
1. A bone age determination method based on artificial intelligence, characterized in that the method comprises:
determining a region of interest in a digitized X-ray DR image, the DR image comprising a skeletal image of a limb;
acquiring a grade evaluation value of the region of interest;
and determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest.
2. The method of claim 1, wherein determining a region of interest in a DR image comprises:
and inputting the DR image into a first neural network model, and detecting the metacarpophalangeal joints in the skeleton image by using the first neural network model to obtain a first region of interest corresponding to the metacarpophalangeal joints.
3. The method of claim 2, wherein prior to determining the region of interest in the DR image, the method further comprises:
generating a first set of training samples for training the first neural network model;
obtaining metacarpophalangeal joint characteristics of each sample in the first training sample set;
and training the first neural network model by taking the metacarpophalangeal joint characteristics as granularity, so that the first neural network model can predict a first region of interest corresponding to the metacarpophalangeal joint.
4. The method of claim 1, wherein determining a region of interest in a DR image comprises:
and inputting the DR image into a second neural network model, and segmenting the wrist joint in the skeleton image by using the second neural network model to obtain a second region of interest corresponding to the wrist joint.
5. The method of claim 4, wherein the segmenting the wrist joint in the bone image by using the second neural network model to obtain a second region of interest corresponding to the wrist joint comprises:
segmenting a wrist joint in the bone image by using the second neural network model, and determining a contour line comprising the wrist joint;
and determining the area in the first shape minimum enclosing frame corresponding to the contour line as the second region of interest.
6. The method of claim 4 or 5, wherein prior to determining a region of interest in the DR image based on the bone image, the method further comprises:
generating a second set of training samples for training the second neural network model;
acquiring wrist joint features of each sample in the second training sample set;
training the second neural network model with the wrist joint features as granularity, so that the second neural network model can predict a second region of interest corresponding to a wrist joint.
7. The method according to claim 1, wherein the obtaining of the rating evaluation value of the region of interest comprises:
inputting the region of interest into a third neural network model, and acquiring a level evaluation value of a metacarpophalangeal joint and/or a wrist joint corresponding to the region of interest by using the third neural network model; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
8. The method according to claim 7, wherein before obtaining the rating evaluation value of the region of interest, the method further comprises:
generating a third set of training samples for training the third neural network model;
obtaining an interested area corresponding to a metacarpophalangeal joint and/or a wrist joint in each sample in the third training sample set;
and training a third neural network model by taking the characteristics of the interested areas corresponding to the metacarpophalangeal joints and/or the wrist joints as granularity, so that the second neural network model can predict the grade evaluation values corresponding to the metacarpophalangeal joints and/or the wrist joints.
9. The method according to claim 1, wherein the determining the bone age corresponding to the bone image based on the level evaluation value of the region of interest comprises:
calculating the sum of the grade evaluation values of all the interested areas;
and determining the bone age corresponding to the bone image based on the sum of the grade evaluation values.
10. A method for training a bone age determination model, the method comprising:
generating a sample set of historical digitized X-ray DR images, the historical DR images including a skeletal image of a limb;
acquiring joint features of each sample in the sample set;
training an interested region neural network model by taking the joint features as granularity, so that the interested region neural network model can predict an interested region corresponding to the joint;
obtaining an interested area corresponding to a joint in each sample;
training a grade evaluation model by taking the characteristics of the interested area of the joint as granularity, so that the grade evaluation model can predict a grade evaluation value corresponding to the joint; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
11. The method of claim 10, wherein the training a region of interest neural network model with the joint feature as a granularity comprises:
training a first neural network model by taking the metacarpophalangeal joint features in the joint features as granularity, so that the first neural network model can predict a first region of interest corresponding to the metacarpophalangeal joint;
and/or training a second neural network model by taking the wrist joint features in the joint features as granularity, so that the second neural network model can predict a second region of interest corresponding to the wrist joint.
12. An artificial intelligence based bone age determination apparatus, comprising:
a region-of-interest determination module for determining a region of interest in a digitized X-ray DR image, the DR image including a skeletal image of a limb;
a grade evaluation value acquisition module, configured to acquire a grade evaluation value of the region of interest;
and the bone age determining module is used for determining the bone age corresponding to the bone image based on the grade evaluation value of the region of interest.
13. A bone age determination model training apparatus, characterized in that the apparatus comprises:
a generation module to generate a historical set of digitized X-ray DR image samples, the historical DR image comprising a skeletal image of a limb;
the joint feature acquisition module is used for acquiring the joint features of each sample in the sample set;
the first training module is used for training an interested region neural network model by taking the joint features as granularity, so that the interested region neural network model can predict an interested region corresponding to a joint;
a region of interest acquisition module for acquiring a region of interest corresponding to a joint in each sample
The second training module is used for training a grade evaluation model by taking the characteristics of the interested area of the joint as granularity, so that the grade evaluation model can predict a grade evaluation value corresponding to the joint; and the grade evaluation value is used for representing the bone age corresponding to the bone image.
14. An electronic device, comprising:
a memory for storing executable instructions;
a processor for implementing the artificial intelligence based bone age determination method of any one of claims 1 to 9 when executing executable instructions stored in the memory;
alternatively, the processor, when executing the executable instructions stored in the memory, is configured to implement the bone age determination model training method of claim 10 or 11.
15. A computer-readable storage medium storing executable instructions for, when executed by a processor, implementing the artificial intelligence based bone age determination method of any one of claims 1 to 9; alternatively, the bone age determination model training method of claim 10 or 11 is implemented.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110566880.9A CN113436145A (en) | 2021-05-24 | 2021-05-24 | Bone age determination method and device based on artificial intelligence and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110566880.9A CN113436145A (en) | 2021-05-24 | 2021-05-24 | Bone age determination method and device based on artificial intelligence and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113436145A true CN113436145A (en) | 2021-09-24 |
Family
ID=77802774
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110566880.9A Pending CN113436145A (en) | 2021-05-24 | 2021-05-24 | Bone age determination method and device based on artificial intelligence and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113436145A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115222955A (en) * | 2022-06-13 | 2022-10-21 | 北京医准智能科技有限公司 | Training method and device of image matching model, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741309A (en) * | 2018-12-27 | 2019-05-10 | 北京深睿博联科技有限责任公司 | A kind of stone age prediction technique and device based on depth Recurrent networks |
| CN110782450A (en) * | 2019-10-31 | 2020-02-11 | 北京推想科技有限公司 | Hand carpal development grade determining method and related equipment |
| CN111415334A (en) * | 2020-03-05 | 2020-07-14 | 北京深睿博联科技有限责任公司 | Bone age prediction device |
| CN112233128A (en) * | 2020-10-15 | 2021-01-15 | 推想医疗科技股份有限公司 | Image segmentation method, model training method, device, medium, and electronic device |
-
2021
- 2021-05-24 CN CN202110566880.9A patent/CN113436145A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741309A (en) * | 2018-12-27 | 2019-05-10 | 北京深睿博联科技有限责任公司 | A kind of stone age prediction technique and device based on depth Recurrent networks |
| CN110782450A (en) * | 2019-10-31 | 2020-02-11 | 北京推想科技有限公司 | Hand carpal development grade determining method and related equipment |
| CN111415334A (en) * | 2020-03-05 | 2020-07-14 | 北京深睿博联科技有限责任公司 | Bone age prediction device |
| CN112233128A (en) * | 2020-10-15 | 2021-01-15 | 推想医疗科技股份有限公司 | Image segmentation method, model training method, device, medium, and electronic device |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115222955A (en) * | 2022-06-13 | 2022-10-21 | 北京医准智能科技有限公司 | Training method and device of image matching model, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Thomas et al. | Automated classification of radiographic knee osteoarthritis severity using deep neural networks | |
| US11423538B2 (en) | Computer-implemented machine learning for detection and statistical analysis of errors by healthcare providers | |
| US20200334416A1 (en) | Computer-implemented natural language understanding of medical reports | |
| US11521716B2 (en) | Computer-implemented detection and statistical analysis of errors by healthcare providers | |
| KR101864380B1 (en) | Surgical image data learning system | |
| CN109741309A (en) | A kind of stone age prediction technique and device based on depth Recurrent networks | |
| CN110827236B (en) | Brain tissue layering method, device and computer equipment based on neural network | |
| KR20190088375A (en) | Surgical image data learning system | |
| JP2021532454A (en) | Knockout autoencoder for detecting anomalies in biomedical images | |
| CN114359970A (en) | Pedestrian re-identification method and device, electronic equipment and storage medium | |
| CN110648318A (en) | Auxiliary analysis method and device for skin diseases, electronic equipment and storage medium | |
| CN113707309A (en) | Disease prediction method and device based on machine learning | |
| Soorya et al. | Automated framework for screening of glaucoma through cloud computing | |
| Vakanski et al. | Metrics for performance evaluation of patient exercises during physical therapy | |
| CN113436145A (en) | Bone age determination method and device based on artificial intelligence and electronic equipment | |
| Esquembri et al. | Methodology for the implementation of real-time image processing systems using FPGAs and GPUs and their integration in EPICS using Nominal Device Support | |
| US20200178918A1 (en) | Standardizing breast density assessments | |
| Kim et al. | Bone age assessment using artificial intelligence in Korean pediatric population: a comparison of deep-learning models trained with healthy chronological and Greulich-Pyle ages as labels | |
| KR20220007030A (en) | Growth analysis prediction apparatus using bone maturity distribution by interest area and method thereof | |
| CN113436144A (en) | Joint rating method and device based on artificial intelligence and electronic equipment | |
| AU2019204365C1 (en) | Method and System for Image Segmentation and Identification | |
| CN113436143A (en) | Joint detection method and device based on artificial intelligence and electronic equipment | |
| CN116453226A (en) | Human body posture recognition method and device based on artificial intelligence and related equipment | |
| CN113705687B (en) | Image instance labeling method based on artificial intelligence and related equipment | |
| Suh et al. | Bone Age Estimation and Prediction of Final Adult Height Using Deep Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210924 |