CN110211674B

CN110211674B - Bone age testing method and related equipment based on machine learning model

Info

Publication number: CN110211674B
Application number: CN201910327391.0A
Authority: CN
Inventors: 杨剑青
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2019-04-23
Filing date: 2019-04-23
Publication date: 2023-06-27
Anticipated expiration: 2039-04-23
Also published as: CN110211674A

Abstract

The invention discloses a bone age testing method, a device, computer equipment and a storage medium based on a machine learning model, which belong to the technical field of machine learning, and the bone age testing method based on the machine learning model comprises the following steps: receiving an X-ray film to be detected, wherein the X-ray film comprises an image of a bone; identifying attributes of a plurality of preset parts of the bone from the image of the bone of the X-ray film to form a first attribute set; inputting the first attribute set into a first machine learning model to output a first bone age; matching the image of the bone in the X-ray film with the standard image of the human bone with each bone age, and finding the bone age of the human bone matched with the image of the bone in the X-ray film as a second bone age; scoring the attribute of each preset part in the first attribute set, obtaining a total score according to the score of the attribute of each preset part, and outputting a third bone age according to the total score; and outputting the weighted sum of the first bone age, the second bone age and the third bone age as the detected bone age. This avoids detection bias caused by a single bone age algorithm.

Description

Bone age testing method and related equipment based on machine learning model

Technical Field

The present invention relates to the field of machine learning technologies, and in particular, to a bone age testing method, apparatus, computer device, and storage medium based on a machine learning model.

Background

The bone age detection system in the current market can only use a single algorithm to detect the bone age of a patient, the algorithm is not in contrast combination with a plurality of bone age detection methods, and larger errors can occur in partial scenes, so that the risk of delaying the diagnosis of the patient exists.

Disclosure of Invention

Based on the method and the device, the computer equipment and the storage medium, the invention provides a bone age testing method and device based on a machine learning model, and aims to solve the technical problem that a bone age testing system in the related art is single in algorithm and poor in reliability.

In a first aspect, a bone age testing method based on a machine learning model is provided, including:

receiving an X-ray film to be detected, wherein the X-ray film comprises an image of a bone;

identifying attributes of a plurality of preset parts of the bone from an image of the bone of the X-ray film, wherein the identified attributes of the plurality of preset parts of the bone form a first attribute set;

inputting the first attribute set into a first machine learning model, and outputting a first bone age by the first machine learning model;

matching the image of the bone in the X-ray film with the standard image of the human bone with each bone age, and finding the bone age of the human bone matched with the image of the bone in the X-ray film as a second bone age;

scoring the attribute of each preset part in the first attribute set, obtaining a total score according to the score of the attribute of each preset part, inquiring a corresponding relation table of the total score and the bone age according to the total score, and outputting a third bone age;

And calculating a weighted sum of the first bone age, the second bone age and the third bone age as a detected bone age output.

In one embodiment, the first machine learning model is trained as follows:

acquiring an X-ray sample set comprising an image of a bone, wherein each X-ray sample in the X-ray sample set is provided with a bone age label in advance, identifying the attribute of a plurality of preset parts of the bone from each X-ray sample in the X-ray sample set, inputting the identified attribute into a first machine learning model, outputting the judged bone age by the first machine learning model, comparing the judged bone age with the attached bone age label, and if the judged bone age is inconsistent with the attached bone age label, adjusting the first machine learning model to enable the bone age output by the machine learning model to be consistent with the label.

In one embodiment, according to the matching of the image of the bone in the X-ray film with the standard image of the human bone of each bone age, the bone age of the human bone matched with the image of the bone in the X-ray film is found as the second bone age, which specifically includes:

extracting an image of the bone from the X-ray film;

subtracting the pixel value of each pixel of the image of the bone extracted from the X-ray film from the pixel value of the corresponding position of the standard image of the bone of the human with age of each bone and obtaining an absolute value to obtain the absolute value of the pixel value difference of each pixel;

Averaging the absolute values of the pixel value differences of the pixels to obtain an average value;

and taking the bone age corresponding to the standard image with the minimum average value as a second bone age.

In one embodiment, the step of finding the second bone age, which is the bone age of the human bone matching the image of the bone in the X-ray film, according to the matching of the image of the bone in the X-ray film with the standard image of the human bone of each bone age, includes:

inputting the X-ray film into a second machine learning model, and outputting the similarity between the standard image of the human bone of each bone age and the X-ray film by the second machine learning model;

and taking the bone age of the standard image with the minimum similarity as a second bone age.

Wherein the second machine learning model is trained as follows:

inputting each bone image sample pair in the set with the bone image sample pair into a second machine learning model, wherein each bone image sample pair comprises a pair of bone image samples, a similarity label is pre-attached to the pair of bone image samples in advance, the second machine learning model outputs the judged similarity of the bone image sample pair, the judged similarity of the bone image sample pair is compared with the attached label, and if the judged similarity is inconsistent with the attached label, the second machine learning model is adjusted, so that the similarity output by the second machine learning model is consistent with the label.

In one embodiment, scoring the attributes of each predetermined location in the first set of attributes includes:

searching a corresponding relation table of the growth period of the attribute and the bone part according to the attribute of each preset part to obtain the growth period corresponding to the attribute;

searching a corresponding relation table of the growing period and the scoring, thereby finding the scoring corresponding to the growing period.

In one embodiment, deriving the total score from the score for the attribute of each predetermined location includes:

a weighted sum of the scores of the attributes of each predetermined location is calculated as a total score.

In one embodiment, an X-ray film to be measured that receives an image including bone specifically includes:

receiving X-ray films to be detected comprising images of bones in response to operation on a main display screen interface;

transmitting the X-ray film to a main display screen interface;

the method further includes, after calculating a weighted sum of the first bone age, the second bone age, and the third bone age as a detected bone age output:

and sending the bone age reference set to an auxiliary display screen interface.

In a second aspect, there is provided a bone age testing device based on a machine learning model, comprising:

a bone image receiving unit for receiving X-ray film of the image including bone to be measured;

A bone attribute identifying unit for identifying the attributes of a plurality of predetermined parts of the bone from the image of the bone of the X-ray film, wherein the identified attributes of the plurality of predetermined parts of the bone form a first attribute set;

a first bone age output unit for inputting the first attribute set into a first machine learning model, the first machine learning model outputting a first bone age;

the second bone age output unit is used for matching the images of the bones in the X-ray film with the standard images of the human bones with various bone ages, and finding the bone ages of the human bones matched with the images of the bones in the X-ray film as second bone ages;

the third bone age output unit is used for scoring the attribute of each preset position in the first attribute set, obtaining a total score according to the score of the attribute of each preset position, inquiring the corresponding relation table of the total score and the bone age according to the total score, outputting the third bone age, and forming a bone age reference set together with the first bone age and the second bone age for reference in diagnosis.

In a third aspect, a computer device is provided, comprising a memory and a processor, the memory having stored therein computer readable instructions that, when executed by the processor, cause the processor to perform the steps of the machine learning model based bone age testing method described above.

In a fourth aspect, a storage medium storing computer readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the machine learning model based bone age testing method described above is provided.

The technical scheme provided by the embodiment of the disclosure can comprise the following beneficial effects:

the bone age testing method, the device, the computer equipment and the storage medium based on the machine learning model are characterized in that the attributes of a plurality of preset parts of the bone are firstly identified from the image of the bone of the X-ray film to be tested, the attributes of the plurality of preset parts of the identified bone form a first attribute set, and then the first attribute set is input into a first machine learning model, so that the first machine learning model outputs the first bone age; then matching the image of the bone in the X-ray film with the standard image of the human bone with each bone age, and finding the bone age of the human bone matched with the image of the bone in the X-ray film as a second bone age; and finally, scoring the attribute of each preset part in the first attribute set, obtaining a total score according to the score of the attribute of each preset part, inquiring a corresponding relation table of the total score and the bone age according to the total score, and outputting a third bone age, so that three bone age detection results obtained by three different methods are obtained, and finally, weighting and calculating the three bone age detection results to obtain a final result which is the bone age finally detected by the scheme, wherein the calculated bone age detection result avoids detection deviation caused by a single bone age algorithm, and compared with the prior bone age detection result, the bone age detection result is more accurate and has reliability.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

FIG. 1 is an environmental diagram of an implementation of a bone age testing method based on a machine learning model provided in one embodiment.

FIG. 2 is a flow chart illustrating a bone age testing method based on a machine learning model, according to an exemplary embodiment.

Fig. 3 is a flowchart of a specific implementation of step S140 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2.

Fig. 4 is a flowchart of another implementation of step S140 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2.

Fig. 5 is a flowchart showing a specific implementation of step S150 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2.

Fig. 6 is a flowchart of a specific implementation of step S110 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2.

FIG. 7 is a flow chart illustrating another bone age testing method based on a machine learning model according to the corresponding embodiment of FIG. 6.

Fig. 8 is a block diagram illustrating a bone age testing apparatus based on a machine learning model, according to an example embodiment.

FIG. 9 is a block diagram illustrating another bone age testing device based on a machine learning model, according to an example embodiment.

Fig. 10 schematically shows an example block diagram of an electronic device for implementing the agent allocation method described above.

Fig. 11 schematically illustrates a computer-readable storage medium for implementing the agent allocation method described above.

Detailed Description

The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

Fig. 1 is a diagram of an implementation environment of a bone age testing method based on a machine learning model provided in one embodiment, as shown in fig. 1, in which a computer device 100 and a terminal 200 are included.

The computer device 100 is a medical system device, for example, a computer device used by a doctor, a server, or the like. The terminal 200 stores therein an X-ray film including an image of a bone to be measured. The patient may send the X-ray film to the computer device 100 through the terminal 200, and the computer device 100 identifies the attributes of the plurality of predetermined portions of the bone from the image of the bone of the X-ray film, and the identified attributes of the plurality of predetermined portions of the bone form a first attribute set; inputting the first attribute set into a first machine learning model, and outputting a first bone age by the first machine learning model; according to the image of the bone in the X-ray film and the standard image of the human bone with each bone age, the bone age of the human bone matched with the image of the bone in the X-ray film is found to be used as the second bone age; scoring the attribute of each preset part in the first attribute set, obtaining a total score according to the score of the attribute of each preset part, inquiring a corresponding relation table of the total score and the bone age according to the total score, outputting a third bone age, finally, weighting and calculating according to the first bone age, the second bone age and the third bone age, and outputting a final result of calculation as the detected bone age.

It should be noted that, the terminal 200 and the computer device 100 may be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc., but are not limited thereto. The computer device 100 and the terminal 200 may be connected by bluetooth, USB (Universal Serial Bus ) or other communication connection, which is not limited herein.

As shown in fig. 2, in one embodiment, a bone age testing method based on a machine learning model is provided, and the bone age testing method based on the machine learning model may be applied to the computer device 100, and specifically may include the following steps:

step S110, an X-ray film including an image of a bone to be measured is received.

When automatic bone age detection is performed, an X-ray film including an image of a bone to be detected needs to be received first, and then judgment is performed according to the X-ray film. The X-ray film may be directly sent by a hospital X-ray machine, or may be sent by a device with a storage function of the patient, such as a mobile phone, a computer, or a mobile hard disk of the patient, or may be sent by a server of a hospital where the patient has seen a doctor, which is not limited herein.

Step S120, identifying attributes of a plurality of predetermined portions of the bone from the image of the bone of the X-ray film, wherein the identified attributes of the plurality of predetermined portions of the bone form a first attribute set.

Wherein the predetermined portion is, for example, a bone head, a bone joint, a place where the bone middle is the finest, or the like. Such as femoral head diameter, periosteum thickness, etc.

Step S130, inputting the first attribute set into a first machine learning model, and outputting the first bone age by the first machine learning model.

Wherein the first machine learning model is trained as follows:

Since the bone age tag attached to the sample is known, the bone age of the sample is known. The machine learning model is trained with the known results as the desired output. The learning mode is as follows: the connection weight of the network is continuously changed under the stimulation of an external input sample. The essence of learning is to dynamically adjust each connection weight. Since the desired output is known, if the result of the machine learning model output does not match the desired output, the connection weights are automatically adjusted until the resulting output result matches the desired output. In this way, the first machine learning model is trained. When the first machine learning model is trained well enough, the first machine learning model can output a first bone age as long as the first attribute set in the X-ray film to be tested is input into the first machine learning model one by one.

The self-adaptive method is a method based on a machine learning model, and the basic principle is that the bone age is comprehensively judged according to the development conditions of different characteristic positions of bones in an X-ray film through the machine learning model. The basic principle is that the attribute of a plurality of preset parts of the bone in the image of the bone of the X-ray film is extracted, the attribute characteristics of each preset part can show the bone age of the bone, the attribute of the preset parts is input into a first machine learning model, and the first machine learning model comprehensively judges the bone age of the bone according to the attribute characteristics of each preset part.

Step S140, the image of the bone in the X-ray film is matched with the standard image of the human bone with each bone age, and the bone age of the human bone matched with the image of the bone in the X-ray film is found to be used as the second bone age.

The second bone age is the bone age measured by using an atlas method, and the basic principle is that an X-ray film to be measured is compared with a standard atlas to obtain a bone age detection result.

In the method, the standard map which is most similar to the X-ray film to be detected can be obtained by extracting the pixel point contrast of the same position of the X-ray film to be detected and the standard map, and the standard map which is most similar to the X-ray film to be detected can also be obtained by extracting the image contrast of the same position.

Step S150, scoring the attribute of each preset part in the first attribute set, obtaining a total score according to the score of the attribute of each preset part, inquiring a corresponding relation table of the total score and the bone age according to the total score, and outputting a third bone age.

The third bone age is the bone age measured by using a calculation method, and the basic principle is that the bone age detection result is obtained according to the X-ray characteristic standard calculation values of different development periods of bones.

In the invention, the attribute of a plurality of preset parts of the bone in the image of the bone of the X-ray film is extracted, the attribute feature of each preset part can show the development condition of the bone, the preset parts are respectively scored according to the development condition, then all the scores are summed to obtain the total score of the bone, and finally the bone age is determined according to the total score of the bone.

In one embodiment the first, second and third bone ages may together comprise a bone age reference set for reference in diagnosis. Therefore, doctors can obtain the bone age detection result of self diagnosis according to the results of the three algorithms after comprehensively considering, detection deviation caused by a single bone age algorithm is avoided, the bone age detection result has better referential property, and a more accurate result is convenient to obtain.

Step S160, calculating a weighted sum of the first bone age, the second bone age, and the third bone age as the detected bone age.

In one embodiment, the weights of the first bone age, the second bone age, and the third bone age are pre-assigned. For example, when the weight of the first bone age is assigned to be 0.5, the weight of the second bone age is assigned to be 0.2, and the weight of the third bone age is assigned to be 0.3, it can be determined that the bone age is 9.15 when the detection result is that the first bone age is 9 years old, the second bone age is 9 years old, and the third bone age is 9.5 years old.

In another embodiment, the sum of the weights of the first bone age, the second bone age, and the third bone age is 1, and the weights of the first bone age, the second bone age, and the third bone age may be determined by subtracting the sum of the weights of the second bone age and the third bone age from 1, respectively. The second bone age may be determined according to the direct similarity between the X-ray film and the standard map closest thereto, and the specific manner of determination may be that the direct similarity between the X-ray film and the standard map closest thereto is multiplied by 0.5. The third bone age may be determined by determining the error degree of the score, and the specific manner of determining the third bone age may be to subtract the error degree of the score from 1 and then multiply the error degree of the score by 0.5.

The direct similarity between the X-ray film and the standard chart closest thereto may be determined by dividing the absolute value of the difference between the pixel values of the X-ray film and the standard chart closest thereto by 255, or by dividing the absolute value of the difference between the pixel values of the X-ray film and the standard chart closest thereto by the pixel values of the X-ray film, which is not limited herein.

The error degree of the score may be determined by dividing the absolute value of the difference between the score obtained by the score and the median value of the score interval corresponding to the third bone age in the score-to-bone age correspondence table by the score, or may be determined by dividing the score obtained by the score by the quotient of the total score and the maximum value of the score interval corresponding to the third bone age in the score-to-bone age correspondence table and the average value of the quotient of the score obtained by dividing the total score by the maximum value of the score interval corresponding to the third bone age in the score-to-bone age correspondence table.

Different weights are allocated according to the importance degrees of the first bone age, the second bone age and the third bone age. Thus, the means for determining bone age by weighted summation is more comprehensive and the determined bone age is more accurate than the prior art by GP mapping or TW3 detection alone.

Optionally, fig. 3 is a detailed description of step S140 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2, where step S140 may include the following steps:

step S141, extracting bone images from the X-ray film;

step S142, subtracting the pixel value of each pixel of the image of the bone extracted from the X-ray film from the pixel value of the corresponding position of the standard image of the bone of the human of each bone age and obtaining an absolute value to obtain the absolute value of the pixel value difference of each pixel;

step S143, the absolute value of the pixel value difference of each pixel is averaged to obtain an average value;

and step S144, taking the bone age corresponding to the standard image with the minimum average value as a second bone age.

In the atlas method, the most basic principle is to compare the X-ray film to be measured with a standard atlas of one standard atlas every year, and select the bone age of the standard atlas closest to the X-ray film to be measured as the bone age of the X-ray film to be measured.

In this embodiment, the comparison method is to extract the pixel point comparison of the same position of the X-ray film to be detected and the standard map, and obtain the standard map most similar to the X-ray film to be detected. The specific method is that each pixel point on the X-ray film is compared with the pixel value of the pixel point at the same position on the standard map to obtain the absolute value of the difference. For example, at the same position of the pixel point, the pixel value of the X-ray film is 250, and the standard map is 248, the absolute value of the difference is 2, and the difference is smaller, if at the same position of the pixel point, the pixel value of the X-ray film is 250, and the standard map is 48, it is quite possible that a bone grows on the X-ray film at the position, but the bone does not start to develop or develop to the extent that the bone grows on the X-ray film at the corresponding position of the standard map, and the difference is larger. Since the difference between the pixel values can represent the difference between the two images in one dimension, there can be a well-defined standard representation of the difference between the X-ray film and the standard map. And then adding and averaging each absolute value, so that the difference between the whole X-ray film and the standard map can be digitally displayed, and finally, the bone age of the standard map with the minimum average value can be used as the second bone age.

Optionally, fig. 4 is a detailed description of step S140 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2, where step S140 may include the following steps:

step S146, inputting the X-ray film into a second machine learning model, and outputting the similarity between the standard image of the human bone of each bone age and the X-ray film by the second machine learning model;

step S147, taking the bone age of the standard image with the minimum similarity as the second bone age.

Wherein the second machine learning model is trained as follows:

Since the similarity label attached to the pair of samples is known, the similarity of the two images is known. The machine learning model is trained with the known results as the desired output. The learning mode is as follows: the connection weight of the network is continuously changed under the stimulation of an external input sample. The essence of learning is to dynamically adjust each connection weight. Since the desired output is known, if the result of the machine learning model output does not match the desired output, the connection weights are automatically adjusted until the resulting output result matches the desired output. In this way, the second learning model is trained. When the first machine learning model is trained well enough, the second machine learning model outputs the similarity between the standard image of the bone of the person with each bone age and the X-ray film as long as the X-ray films are input into the second machine learning model one by one.

In this embodiment, the similarity between the standard image of the bone of the individual bone age person and the X-ray film is mainly determined by means of a machine learning model, and the determination method of machine learning may be based on the difference between each pixel point or the difference between each comparison block, for example, 3×3 blocks obtained by dividing the standard image and the X-ray film on average.

The method can also be realized through a machine learning model, and the real result of the bone age can be accurately judged through the comparison of the machine learning model with a standard map, so that the subsequent auxiliary diagnosis is facilitated.

Optionally, fig. 5 is a detailed description of step S150 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2, where step S150 may include the following steps:

step S151, searching a corresponding relation table of the growth period of the attribute and the bone part according to the attribute of each preset part to obtain the growth period corresponding to the attribute;

step S152, searching a corresponding relation table of the growing period and the score, thereby finding the score corresponding to the growing period.

For example, the periosteum thickness of the bone part is an attribute, for example, the thickness is 0.6mm, a corresponding relation table of the attribute and the growing period to which the bone part belongs is searched to obtain that the thickness belongs to the growing period 2, and the corresponding score of the growing period 2 is 60 points when the corresponding relation table of the growing period and the score is searched to obtain the corresponding score of 60 points.

Optionally, deriving the total score from the score for the attribute of each of the predetermined locations includes:

The weights of the weighted sums may be set according to specific situations, and different scoring methods may be used to cause weight differences for different human body parts, for example, in one embodiment, when the weight of the pisiform part is set to 0.2, the weight of all phalangeal parts is set to 0.5, the weight of the ulna-radius part is set to 0.2, and the weights of other parts are set to 0.1 for total score calculation.

Thus, the weighted score can be used for comprehensively and accurately judging the bone age of the bone on the X-ray film to be detected, and the deviation of the score is avoided.

Optionally, fig. 6 is a detailed description of step S110 in the bone age testing method based on the machine learning model according to the corresponding embodiment of fig. 2, where step S110 may further include the following steps:

Step S111, receiving an X-ray film to be detected comprising an image of a bone in response to an operation on a main display screen interface;

and step S112, transmitting the X-ray film to a main display screen interface.

In this embodiment, the method enables a doctor to perform a detection operation relatively conveniently by being compatible with the existing PACS system of the hospital, specifically, the method is that the computer device 100 starts to perform a bone age detection by the operation of the doctor, and the computer device 100 sends the X-ray film to a computer main screen of the doctor while receiving the X-ray film, so that the doctor can judge whether the detection result is correct according to the X-ray film and modify the detection result.

Thus, as shown in fig. 7, after the X-ray film is sent to the main display screen interface, the method further includes, after step S160:

and step S180, the bone age reference set is sent to an auxiliary display screen interface.

Therefore, the first bone age, the second bone age and the third bone age are displayed to the doctor, and meanwhile, the doctor is not prevented from observing the details of the bones displayed on the X-ray film to be measured, and the doctor is more facilitated to accurately judge.

The PACS system is compatible with the existing system of the hospital, the X-ray film can be conveniently and directly sent to the bone age detection system, and the calculated result is sent to the auxiliary display screen, so that a doctor can judge the bone age, the bone age detection can be more conveniently and conveniently carried out, and the trouble of using the mobile storage device is avoided.

As shown in fig. 8, in one embodiment, a bone age testing apparatus based on a machine learning model is provided, which may be integrated into the computer device 100 described above, and may specifically include: bone image receiving unit 110, bone attribute identifying unit 120, first bone age output unit 130, second bone age output unit 140, and third bone age output unit 150.

A bone image receiving unit 110 for receiving an X-ray film including an image of a bone to be measured;

a bone attribute identifying unit 120 for identifying attributes of a plurality of predetermined portions of the bone from the image of the bone of the X-ray film, the identified attributes of the plurality of predetermined portions of the bone constituting a first attribute set;

a first bone age output unit 130 for inputting the first set of attributes into a first machine learning model, the first machine learning model outputting a first bone age;

a second bone age output unit 140, configured to find, as a second bone age, a bone age of a human bone that matches the image of the bone in the X-ray film according to matching the image of the bone in the X-ray film with a standard image of the human bone of each bone age;

and a third bone age output unit 150, configured to score the attribute of each predetermined part in the first attribute set, obtain a total score according to the score of the attribute of each predetermined part, query a table of correspondence between the total score and the bone ages according to the total score, output a third bone age, and form a bone age reference set together with the first bone age and the second bone age for reference during diagnosis.

The implementation process of the functions and actions of each module in the device is specifically shown in the implementation process of corresponding steps in the bone age testing method based on the machine learning model, and is not repeated here.

Optionally, fig. 9 is a block diagram of another policy distribution device according to the corresponding embodiment of fig. 8, and as shown in fig. 9, the bone age testing device based on the machine learning model shown in fig. 8 further includes, but is not limited to: a final bone age determination unit 160.

The final bone age determination unit 160 is configured to calculate a weighted sum of the first bone age, the second bone age, and the third bone age as the detected bone age.

It should be noted that although in the above detailed description several modules or units of a device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functionality of two or more modules or units described above may be embodied in one module or unit in accordance with embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into a plurality of modules or units to be embodied.

Furthermore, although the steps of the methods in the present disclosure are depicted in a particular order in the drawings, this does not require or imply that the steps must be performed in that particular order or that all illustrated steps be performed in order to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step to perform, and/or one step decomposed into multiple steps to perform, etc.

From the above description of embodiments, those skilled in the art will readily appreciate that the example embodiments described herein may be implemented in software, or may be implemented in software in combination with the necessary hardware. Thus, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a U-disk, a mobile hard disk, etc.) or on a network, including several instructions to cause a computing device (may be a personal computer, a server, a mobile terminal, or a network device, etc.) to perform the method according to the embodiments of the present disclosure.

In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

Those skilled in the art will appreciate that the various aspects of the invention may be implemented as a system, method, or program product. Accordingly, aspects of the invention may be embodied in the following forms, namely: an entirely hardware embodiment, an entirely software embodiment (including firmware, micro-code, etc.) or an embodiment combining hardware and software aspects may be referred to herein as a "circuit," module "or" system.

An electronic device 500 according to this embodiment of the present invention is described below with reference to fig. 10. The electronic device 500 shown in fig. 10 is merely an example, and should not be construed as limiting the functionality and scope of use of embodiments of the present invention.

As shown in fig. 10, the electronic device 500 is embodied in the form of a general purpose computing device. The components of electronic device 500 may include, but are not limited to: the at least one processing unit 510, the at least one memory unit 520, and a bus 530 connecting the various system components, including the memory unit 520 and the processing unit 510.

Wherein the storage unit stores program code that is executable by the processing unit 510 such that the processing unit 510 performs steps according to various exemplary embodiments of the present invention described in the above section of the "exemplary method" of the present specification. For example, the processing unit 510 may perform step S110 shown in fig. 2, and receive an X-ray film including an image of a bone to be measured; step S120, identifying the attributes of a plurality of preset parts of the bone from the image of the bone of the X-ray film, wherein the identified attributes of the plurality of preset parts of the bone form a first attribute set; step S130, inputting a first attribute set into a first machine learning model, and outputting a first bone age by the first machine learning model; step S140, according to the image of the bone in the X-ray film and the standard image of the human bone with each bone age, the bone age of the human bone matched with the image of the bone in the X-ray film is found to be used as the second bone age; step S150, scoring the attribute of each preset part in the first attribute set, obtaining a total score according to the score of the attribute of each preset part, inquiring a corresponding relation table of the total score and the bone age according to the total score, and outputting a third bone age; step S160, calculating a weighted sum of the first bone age, the second bone age, and the third bone age as the detected bone age.

The storage unit 520 may include readable media in the form of volatile storage units, such as Random Access Memory (RAM) 5201 and/or cache memory unit 5202, and may further include Read Only Memory (ROM) 5203.

The storage unit 520 may also include a program/utility 5204 having a set (at least one) of program modules 5205, such program modules 5205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment.

Bus 530 may be one or more of several types of bus structures including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

The electronic device 500 may also communicate with one or more external devices 700 (e.g., keyboard, pointing device, bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 500, and/or any device (e.g., router, modem, etc.) that enables the electronic device 500 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 550. Also, electronic device 500 may communicate with one or more networks such as a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the Internet, through network adapter 560. As shown, network adapter 560 communicates with other modules of electronic device 500 over bus 530. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with electronic device 500, including, but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.

From the above description of embodiments, those skilled in the art will readily appreciate that the example embodiments described herein may be implemented in software, or may be implemented in software in combination with the necessary hardware. Thus, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a U-disk, a mobile hard disk, etc.) or on a network, including several instructions to cause a computing device (may be a personal computer, a server, a terminal device, or a network device, etc.) to perform the method according to the embodiments of the present disclosure.

In an exemplary embodiment of the present disclosure, a computer-readable storage medium having stored thereon a program product capable of implementing the method described above in the present specification is also provided. In some possible embodiments, the various aspects of the invention may also be implemented in the form of a program product comprising program code for causing a terminal device to carry out the steps according to the various exemplary embodiments of the invention as described in the "exemplary methods" section of this specification, when said program product is run on the terminal device.

Referring to fig. 11, a program product 600 for implementing the above-described method according to an embodiment of the present invention is described, which may employ a portable compact disc read only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium would include the following: an electrical connection having one or more wires, a portable disk, a hard disk, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer readable signal medium may include a data signal propagated in baseband or as part of a carrier wave with readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).

Furthermore, the above-described drawings are only schematic illustrations of processes included in the method according to the exemplary embodiment of the present invention, and are not intended to be limiting. It will be readily appreciated that the processes shown in the above figures do not indicate or limit the temporal order of these processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, for example, among a plurality of modules.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

Claims

1. A method for testing bone age based on a machine learning model, the method comprising:

calculating a weighted sum of the first bone age, the second bone age, and the third bone age as a detected bone age output;

wherein, match the image of bone in the X-ray film with the standard image of the human bone of each bone age, find the bone age of the human bone that matches with the image of bone in the X-ray film as the second bone age, include: extracting an image of the bone from the X-ray film; subtracting the pixel value of each pixel of the image of the bone extracted from the X-ray film from the pixel value of the corresponding position of the standard image of the bone of the human with age of each bone and obtaining an absolute value to obtain the absolute value of the pixel value difference of each pixel; averaging the absolute values of the pixel value differences of the pixels to obtain an average value; taking the bone age corresponding to the standard image with the minimum average value as a second bone age;

Or, matching the image of the bone in the X-ray film with the standard image of the human bone with each bone age, and finding the bone age of the human bone matched with the image of the bone in the X-ray film as a second bone age, wherein the method comprises the following steps: inputting the X-ray film into a second machine learning model, and outputting the similarity between the standard image of the human bone of each bone age and the X-ray film by the second machine learning model; taking the bone age of the standard image with the minimum similarity as a second bone age; wherein the second machine learning model is trained as follows: inputting each bone image sample pair in the set with the bone image sample pair into a second machine learning model, wherein each bone image sample pair comprises a pair of bone image samples, a similarity label is pre-attached to the pair of bone image samples in advance, the second machine learning model outputs the judged similarity of the bone image sample pair, the judged similarity of the bone image sample pair is compared with the attached label, and if the judged similarity is inconsistent with the attached label, the second machine learning model is adjusted, so that the similarity output by the second machine learning model is consistent with the label.

2. The method of claim 1, wherein the first machine learning model is trained as follows:

acquiring an X-ray sample set comprising an image of a bone, wherein each X-ray sample in the X-ray sample set is provided with a bone age label in advance, identifying the attribute of a plurality of preset parts of the bone from each X-ray sample in the X-ray sample set, inputting the identified attribute into a first machine learning model, outputting the judged bone age by the first machine learning model, comparing the judged bone age with the attached bone age label, and if the judged bone age is inconsistent with the attached bone age label, adjusting the first machine learning model until the bone age output by the machine learning model is consistent with the label.

3. The method of claim 1, wherein scoring the attribute of each predetermined location in the first set of attributes comprises:

4. The method of claim 1, wherein deriving a total score based on the score for the attribute of each predetermined location comprises:

5. The method of claim 1 or 2, wherein receiving the X-ray film to be measured comprising an image of the bone, comprises:

transmitting the X-ray film to a main display screen interface;

and sending the bone age reference set to an auxiliary display screen interface, wherein the bone age reference set comprises a first bone age, a second bone age and a third bone age.

6. A bone age testing device based on a machine learning model, the device comprising:

a third bone age output unit, configured to score the attribute of each predetermined part in the first attribute set, obtain a total score according to the score of the attribute of each predetermined part, query a table of correspondence between the total score and the bone ages according to the total score, output a third bone age, and form a bone age reference set together with the first bone age and the second bone age for reference during diagnosis;

wherein, according to the image of bone in the X-ray film and the standard image of human bone of each bone age match, find the bone age of human bone that matches with the image of bone in the X-ray film as second bone age, include: extracting an image of the bone from the X-ray film; subtracting the pixel value of each pixel of the image of the bone extracted from the X-ray film from the pixel value of the corresponding position of the standard image of the bone of the human with age of each bone and obtaining an absolute value to obtain the absolute value of the pixel value difference of each pixel; averaging the absolute values of the pixel value differences of the pixels to obtain an average value; taking the bone age corresponding to the standard image with the minimum average value as a second bone age;

Or, according to the matching of the image of the bone in the X-ray film and the standard image of the human bone with each bone age, finding the bone age of the human bone matched with the image of the bone in the X-ray film as a second bone age, including: inputting the X-ray film into a second machine learning model, and outputting the similarity between the standard image of the human bone of each bone age and the X-ray film by the second machine learning model; taking the bone age of the standard image with the minimum similarity as a second bone age; wherein the second machine learning model is trained as follows: inputting each bone image sample pair in the set with the bone image sample pair into a second machine learning model, wherein each bone image sample pair comprises a pair of bone image samples, a similarity label is pre-attached to the pair of bone image samples in advance, the second machine learning model outputs the judged similarity of the bone image sample pair, the judged similarity of the bone image sample pair is compared with the attached label, and if the judged similarity is inconsistent with the attached label, the second machine learning model is adjusted, so that the similarity output by the second machine learning model is consistent with the label.

7. A computer device comprising a memory and a processor, the memory having stored therein computer readable instructions that, when executed by the processor, cause the processor to perform the method of any of claims 1 to 5.

8. A storage medium storing computer readable instructions which, when executed by one or more processors, cause the one or more processors to perform the method of any of claims 1 to 5.