CN112102942B - Skeletal development grade detection method and terminal equipment - Google Patents

Skeletal development grade detection method and terminal equipment Download PDF

Info

Publication number
CN112102942B
CN112102942B CN202011015997.XA CN202011015997A CN112102942B CN 112102942 B CN112102942 B CN 112102942B CN 202011015997 A CN202011015997 A CN 202011015997A CN 112102942 B CN112102942 B CN 112102942B
Authority
CN
China
Prior art keywords
level
grade
development
bone
preset number
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202011015997.XA
Other languages
Chinese (zh)
Other versions
CN112102942A (en
Inventor
张淼
李文旭
庞海
张邵岩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shijiazhuang Hi Tech Co ltd
Original Assignee
Shijiazhuang Hi Tech Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shijiazhuang Hi Tech Co ltd filed Critical Shijiazhuang Hi Tech Co ltd
Priority to CN202011015997.XA priority Critical patent/CN112102942B/en
Publication of CN112102942A publication Critical patent/CN112102942A/en
Application granted granted Critical
Publication of CN112102942B publication Critical patent/CN112102942B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic 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
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Molecular Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Probability & Statistics with Applications (AREA)
  • Evolutionary Biology (AREA)
  • Databases & Information Systems (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Pathology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Image Analysis (AREA)

Abstract

The invention is applicable to the technical field of bone age, and discloses a bone development grade detection method and terminal equipment, wherein the method comprises the following steps: constructing a skeleton development grade detection model, wherein an output layer of the skeleton development grade detection model comprises a first output branch and a second output branch, the first output branch adopts a softmax logistic regression method to carry out grade classification, and the second output branch adopts a linear regression method to carry out grade classification; training the skeletal development level detection model to obtain a trained skeletal development level detection model; acquiring a bone image, inputting the bone image into a trained bone development grade detection model, and obtaining a first classification result output by a first output branch and a second classification result output by a second output branch; and determining the development grade of the skeleton corresponding to the skeleton image according to the first classification result and the second classification result. The invention can improve the accuracy and stability of skeletal development level detection.

Description

Skeletal development grade detection method and terminal equipment
Technical Field
The invention belongs to the technical field of bone age, and particularly relates to a bone development level detection method and terminal equipment.
Background
In the field of bone age, classification of bone development grades has the specificity which is not possessed by natural image classification, the growth and development process of organisms needs to be considered to be a linear process, the boundaries between different grades of bones are less obvious than natural images, and the grades have clear sequence.
At present, a softmax logistic regression model is generally adopted to detect the development grade of bones, but the softmax logistic regression model considers mutual exclusivity among different categories, excludes the relevance of the different categories, does not consider that the growth and development process of organisms is a linear process, and leads to discrete development grade results of the output bones and easy occurrence of larger deviation.
Disclosure of Invention
In view of the above, the embodiment of the invention provides a skeletal development level detection method and terminal equipment, so as to solve the problem that a skeletal development level result output by the prior art is easy to deviate greatly.
A first aspect of an embodiment of the present invention provides a skeletal development level detection method, including:
constructing a skeleton development grade detection model, wherein an output layer of the skeleton development grade detection model comprises a first output branch and a second output branch, the first output branch adopts a softmax logistic regression method to carry out grade classification, and the second output branch adopts a linear regression method to carry out grade classification;
training the skeletal development level detection model to obtain a trained skeletal development level detection model;
acquiring a bone image, inputting the bone image into a trained bone development grade detection model, and obtaining a first classification result output by a first output branch and a second classification result output by a second output branch;
and determining the development grade of the skeleton corresponding to the skeleton image according to the first classification result and the second classification result.
A second aspect of an embodiment of the present invention provides a terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the skeletal development level detection method of the first aspect when the computer program is executed by the processor.
Compared with the prior art, the embodiment of the invention has the beneficial effects that: the output layer of the skeleton development grade detection model constructed by the embodiment of the invention comprises a first output branch and a second output branch, wherein the first output branch adopts a softmax logistic regression method for grade classification, the second output branch adopts a linear regression method for grade classification, a first classification result output by the first output branch and a second classification result output by the second output branch are obtained by inputting skeleton images into the trained skeleton development grade detection model, and the development grade of skeleton corresponding to the skeleton images is determined according to the first classification result and the second classification result.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a bone development level detection method according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of an implementation of a skeletal development level detection method according to an embodiment of the present invention;
FIG. 3 is a schematic block diagram of a skeletal development level detection system provided in one embodiment of the present invention;
fig. 4 is a schematic block diagram of a terminal device according to an embodiment of the present invention.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
In order to illustrate the technical scheme of the invention, the following description is made by specific examples.
Fig. 1 is a schematic flow chart of an implementation of a skeletal development level detection method according to an embodiment of the present invention, and for convenience of explanation, only the portions related to the embodiment of the present invention are shown. The execution body of the embodiment of the invention can be a terminal device. As shown in fig. 1, the method may include the steps of:
s101: and constructing a skeletal development grade detection model, wherein an output layer of the skeletal development grade detection model comprises a first output branch and a second output branch, the first output branch adopts a softmax logistic regression method to carry out grade classification, and the second output branch adopts a linear regression method to carry out grade classification.
Optionally, "constructing a skeletal development level detection model" in S101 above may include:
constructing a basic convolutional neural network model;
and constructing a first output branch and a second output branch at an output layer of the basic convolutional neural network model to obtain a skeleton development level detection model.
The basic convolutional neural network model may be any convolutional neural network model capable of realizing functions such as feature extraction, and the like, and is not particularly limited herein.
According to the embodiment of the invention, two output branches are constructed at the output layer of the basic convolutional neural network model, and the softmax logistic regression method and the linear regression (Linear Regreesion) method are respectively adopted for classification, so that the advantages of the softmax logistic regression classification and the continuity of the linear regression can be considered, and the skeleton development grade detection method is more suitable for the continuous classification field of natural growth development and is suitable for skeleton development grade detection.
S102: training the skeletal development level detection model to obtain a trained skeletal development level detection model.
Optionally, in training the skeletal development level detection model, the loss function of the first output branch is cross entropy and the loss function of the second output branch is mean square error.
In the embodiment of the invention, the skeletal development level detection model can be trained by adopting the existing method based on a preset training sample set, so as to obtain the trained skeletal development level detection model. Wherein during training, the first output branch adopts cross entropy (cross-entropy) as a loss function, and the second output branch adopts Mean Square Error (MSE) as a loss function.
S103: and acquiring a bone image, inputting the bone image into a trained bone development level detection model, and obtaining a first classification result output by the first output branch and a second classification result output by the second output branch.
Wherein the bone image is an image of the bone of which the developmental level is to be detected.
The first output branch may output a plurality of classification results and confidence levels respectively corresponding to the plurality of classification results, each classification result being a development level, the confidence levels representing probabilities that bone development levels are classification results corresponding to the confidence levels.
The second output branch may output a predicted value, which is a floating point number, from which the bone development level may be estimated.
S104: and determining the development grade of the skeleton corresponding to the skeleton image according to the first classification result and the second classification result.
In one embodiment of the present invention, the first classification result includes a first preset number of classification results and confidence degrees corresponding to the first preset number of classification results, where the first preset number of classification results are the classification results of the first preset number arranged in the order from the high confidence degree to the low confidence degree among all the classification results output by the first output branch; the second classification result is a floating point number.
Referring to fig. 2, the step S104 may include the steps of:
s201: the floating point number is converted into a corresponding development grade and is marked as a first grade.
In one embodiment of the present invention, the step S201 may include the steps of:
sequentially converting each development grade of the skeleton into a natural number starting from 0 according to the sequence from small to large to obtain a natural number grade sequence;
and determining the closest target natural number of the floating point number from the natural number level sequence, and marking the development level corresponding to the target natural number as a first level.
Illustratively, if the bone age criteria is the RUS-CHN bone age criteria, the total is divided into 15 classes, including classes 0,1,2,3,4,5-0,5-2,6,7-0,7-2,8-0,8-1,8-2,8-3, 8-4. Wherein, for example, 5-2 is a level with a suffix, and the suffix is 2;4 is a no suffix level. The 15 levels are sequentially converted into natural numbers starting from 0 in order from large to small, namely from 0, so that a natural number level sequence is obtained, namely the natural number level sequence is 0,1,2,3,4,5,6,7,8,9, 10, 11, 12, 13 and 14, and each natural number corresponds to one level of the 15 levels in order. If the floating point number is 5.20254, the closest natural number is 5, and therefore the target natural number is 5, and the rank 5-0 corresponding to the natural number 5 is the first rank.
If the bone age standard is TW3 bone age standard, the bone age standard is divided into 9 grades, including 0,1,2,3,4,5,6,7,8 grades, and since the grades of the standard are all natural numbers, the step of converting may be omitted, and the 9 grades are regarded as a natural number grade sequence. Similarly, if the floating point number is 5.20254, the closest natural number is 5, and therefore the target natural number is 5, and rank 5 corresponding to natural number 5 is the first rank.
Because the RUS-CHN bone age standard is compatible with the TW3 bone age standard, the RUS-CHN bone age standard is adopted to train the bone development grade detection model in the embodiment of the invention, and the classification result is the RUS-CHN bone age standard, so that the compatibility is higher.
S202: and selecting the classification result with the highest confidence from the first preset number of classification results, and marking the classification result as a second grade.
The first preset number may be set according to actual requirements, and optionally, the first preset number may be 3 or 2.
S203: and if the first grade is the same as the second grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade.
S204: if the first level is different from the second level, judging whether the second level is reliable or not, obtaining a reliability judging result, and determining the development level of the skeleton corresponding to the skeleton image according to the reliability judging result.
In one embodiment of the present invention, the step of "determining whether the second level is reliable or not to obtain the reliability determination result" in S204 may include the following steps:
determining whether the second level has a suffix;
if the second level has no suffix, executing a first operation based on the first preset number of classification results and the confidence level of the second level to obtain a reliability judgment result;
if the second level has a suffix, determining a target bone age standard;
if the target bone age standard is RUS-CHN bone age standard, executing a first operation based on the first preset number of classification results and the second level of confidence coefficient to obtain a reliability judgment result;
and if the target bone age standard is the bone age standard without the suffix, executing a second operation based on the classification results of the first preset number to obtain a reliability judgment result.
The target bone age standard is a standard of bone development level which needs to be output by a user, and can be RUS-CHN bone age standard or bone age standard without suffix. The levels of the RUS-CHN bone age standard without the suffix and the levels of the bone age standard without the suffix can be commonly used, so that whether the second level has the suffix or not is judged and determined first, and if the second level does not have the suffix, the same method can be adopted for processing no matter which target bone age standard is. If the second level has a suffix, the levels of the two level standards are not universal and are processed by different methods.
Among other things, the bone age criteria without suffix may include the TW3 bone age criteria, the TW2 bone age criteria, the She Shi bone age criteria, and so forth.
In an embodiment of the present invention, the performing a first operation to obtain a reliability determination result based on the first preset number of classification results and the second level of confidence includes:
determining whether the second level of confidence is greater than a confidence threshold;
if the confidence coefficient of the second level is larger than the confidence coefficient threshold value, the reliability judgment result is that the second level is reliable;
if the confidence coefficient of the second level is not greater than the confidence coefficient threshold value, determining whether the levels of the classification results of the first preset number are continuous or not;
if the first preset number of classification results are continuous in level, the reliability judgment result is the second level reliability;
if the grades of the classification results of the first preset number are discontinuous, the reliability judgment result is that the second grade is unreliable.
The confidence coefficient threshold value is obtained empirically, and if the confidence coefficient of the second level is larger than the confidence coefficient threshold value, the confidence coefficient of the second level is higher, more accurate and reliable. If the confidence level of the second level is not high enough, further, determining whether the level of the first preset number of classification results is continuous, for example, the first preset number may be 2, that is, the first preset number of classification results are two levels with the highest confidence level and the second highest confidence level in the plurality of classification results, if the two levels are continuous, the overall trend of the classification results is consistent, the reliability of the second level is higher, and the second level is determined to be reliable; if the two grades are discontinuous, the overall trend of the classification result is inconsistent, the reliability of the second grade is not high, and the second grade is determined to be unreliable.
In an embodiment of the present invention, the performing a second operation based on the first preset number of classification results to obtain a reliability determination result includes:
determining whether the grades of the first preset number of classification results are continuous;
if the first preset number of classification results are continuous in level, the reliability judgment result is the second level reliability;
if the grades of the classification results of the first preset number are discontinuous, the reliability judgment result is that the second grade is unreliable.
In one embodiment of the present invention, determining a bone development level corresponding to a bone image according to a reliability determination result includes:
if the reliability judging result is that the second level is reliable, judging whether the second level accords with the relevance rule, obtaining a relevance judging result, and determining the development level of the skeleton corresponding to the skeleton image according to the relevance judging result;
if the reliability judging result is that the second level is unreliable, determining whether the first preset number of classification results contain the first level or not;
if the first preset number of classification results comprise a first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
if the first classification results of the first preset number do not contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the second grade.
If the reliability judging result is that the second level is unreliable and the first level is included in the first preset number of classification results, determining that the development level of the skeleton corresponding to the skeleton image is the first level, prompting the user that the result is not completely reliable, and giving an unreliable reason, for example, that the second level is unreliable, that the second level is different from the first level, and the like.
If the reliability judging result is that the second level is unreliable and the first level is not included in the first preset number of classification results, determining that the development level of the skeleton corresponding to the skeleton image is the second level, prompting the user that the result is not completely reliable, and giving an unreliable reason, for example, the second level is unreliable, the second level is different from the first level, the first level is not included in the first preset number of classification results, and the like.
In one embodiment of the present invention, determining a bone development level corresponding to a bone image according to a correlation determination result includes:
if the correlation judgment result shows that the second level accords with the correlation rule, determining that the development level of the skeleton corresponding to the skeleton image is the second level;
if the correlation judgment result is that the second level does not accord with the correlation rule, determining whether the first level is contained in the first preset number of classification results;
if the first preset number of classification results comprise a first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
if the first classification results of the first preset number do not contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the second grade.
If the correlation judgment result is that the second level does not accord with the correlation rule and the first level is included in the first preset number of classification results, determining that the development level of the skeleton corresponding to the skeleton image is the first level, prompting the user that the result is not completely reliable and giving an unreliable reason, for example, that the second level does not accord with the correlation rule, that the second level is different from the first level, and the like.
If the correlation judgment result is that the second level does not accord with the correlation rule and the first level is not included in the first preset number of classification results, determining that the development level of the skeleton corresponding to the skeleton image is the second level, prompting the user that the result is not completely reliable, and giving an unreliable reason, for example, that the second level does not accord with the correlation rule, the second level is different from the first level, the first level is not included in the first preset number of classification results, and the like.
Optionally, the determining whether the second level meets the association rule, to obtain an association determination result includes:
determining a second preset number of bones related to bones corresponding to the bone image based on an Aprior algorithm, wherein the second preset number of bones are bones which are arranged in the front second preset number in the order of the support degree from the large to the small in all bones related to bones corresponding to the bone image;
acquiring development grades corresponding to the bones of the second preset number respectively, and calculating the confidence coefficient of the second grade to the development grades corresponding to the bones of the second preset number respectively;
if the confidence coefficient of the second level to the development levels respectively corresponding to the bones of the second preset number is larger than or equal to the preset confidence coefficient, the relevance judgment result is that the second level accords with the relevance rule;
if the confidence degree of the second level on the development levels respectively corresponding to the bones of the second preset number is smaller than the preset confidence degree, the relevance judgment result is that the second level does not accord with the relevance rule.
The preset confidence level and the second preset number can be set according to actual requirements. Alternatively, the second preset number may be 3.
The Aprior algorithm is a commonly used algorithm for mining out data association rules. The algorithm can mine the relevance among different bones, and select the bones with the highest relevance (support degree) from bones related to bones to be detected by a second preset number of bones. Calculating the confidence coefficient of the second level to the development levels respectively corresponding to the bones of the second preset number based on an Aprior algorithm, and judging that the second level accords with the relevance rule if the confidence coefficient is larger than or equal to the preset confidence coefficient; if the confidence is smaller than the preset confidence, the second level is judged to be not in accordance with the association rule.
As can be seen from the above description, the output layer of the bone development level detection model constructed according to the embodiment of the present invention includes a first output branch and a second output branch, where the first output branch performs level classification by using a softmax logistic regression method, and the second output branch performs level classification by using a linear regression method, and the first classification result output by the first output branch and the second classification result output by the second output branch are obtained by inputting the bone image into the trained bone development level detection model, and the development level of the bone corresponding to the bone image is determined according to the first classification result and the second classification result, so that the bone development level detection method according to the embodiment of the present invention is more suitable for the continuous classification field of natural growth development and can improve the accuracy of bone development level detection. Because the result of softmax logistic regression classification is discrete, the result of comparison jump classification can be possibly caused, and the constraint model can be better through the constraint of the linear regression branch, so that the result of jump larger cannot occur to the classification branch, and the stability of the skeleton development level detection result can be ensured.
It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present invention.
Corresponding to the above bone development level detection method, an embodiment of the present invention further provides a bone development level detection system, which has the same beneficial effects as the above bone development level detection method. Fig. 3 is a schematic block diagram of a skeletal development level detection system provided in an embodiment of the present invention, and for ease of illustration only parts relevant to the embodiment of the present invention are shown.
In an embodiment of the present invention, skeletal development level detection system 30 may include a model building module 301, a training module 302, a classification module 303, and a development level determination module 304.
The building module 301 is configured to build a skeletal development level detection model, where an output layer of the skeletal development level detection model includes a first output branch and a second output branch, the first output branch performs level classification by using a softmax logistic regression method, and the second output branch performs level classification by using a linear regression method;
the training module 302 is configured to train the skeletal development level detection model to obtain a trained skeletal development level detection model;
the classification module 303 is configured to obtain a bone image, and input the bone image into a trained bone development level detection model to obtain a first classification result output by the first output branch and a second classification result output by the second output branch;
the development level determining module 304 is configured to determine a development level of a bone corresponding to the bone image according to the first classification result and the second classification result.
Optionally, the first classification result includes a first preset number of classification results and confidence degrees corresponding to the first preset number of classification results, where the first preset number of classification results are the classification results arranged in the first preset number according to the order of the confidence degrees from the high to the low in all the classification results output by the first output branch; the second classification result is a floating point number;
the development level determining module 304 is specifically configured to:
converting the floating point number into a corresponding development grade, and marking the development grade as a first grade;
selecting the classification result with the highest confidence from the first preset number of classification results, and marking the classification result as a second grade;
if the first grade is the same as the second grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
if the first level is different from the second level, judging whether the second level is reliable or not, obtaining a reliability judging result, and determining the development level of the skeleton corresponding to the skeleton image according to the reliability judging result.
Optionally, the developmental grade determination module 304 may also be configured to:
sequentially converting each development grade of the skeleton into a natural number starting from 0 according to the sequence from small to large to obtain a natural number grade sequence;
and determining the closest target natural number of the floating point number from the natural number level sequence, and marking the development level corresponding to the target natural number as a first level.
Optionally, the developmental grade determination module 304 may also be configured to:
determining whether the second level has a suffix;
if the second level has no suffix, executing a first operation based on the first preset number of classification results and the confidence level of the second level to obtain a reliability judgment result;
if the second level has a suffix, determining a target bone age standard;
if the target bone age standard is RUS-CHN bone age standard, executing a first operation based on the first preset number of classification results and the second level of confidence coefficient to obtain a reliability judgment result;
and if the target bone age standard is the bone age standard without the suffix, executing a second operation based on the classification results of the first preset number to obtain a reliability judgment result.
Optionally, the developmental grade determination module 304 may also be configured to:
determining whether the second level of confidence is greater than a confidence threshold;
if the confidence coefficient of the second level is larger than the confidence coefficient threshold value, the reliability judgment result is that the second level is reliable;
if the confidence coefficient of the second level is not greater than the confidence coefficient threshold value, determining whether the levels of the classification results of the first preset number are continuous or not;
if the first preset number of classification results are continuous in level, the reliability judgment result is the second level reliability;
if the grades of the classification results of the first preset number are discontinuous, the reliability judgment result is that the second grade is unreliable.
Optionally, the developmental grade determination module 304 may also be configured to:
determining whether the grades of the first preset number of classification results are continuous;
if the first preset number of classification results are continuous in level, the reliability judgment result is the second level reliability;
if the grades of the classification results of the first preset number are discontinuous, the reliability judgment result is that the second grade is unreliable.
Optionally, the developmental grade determination module 304 may also be configured to:
if the reliability judging result is that the second level is reliable, judging whether the second level accords with the relevance rule, obtaining a relevance judging result, and determining the development level of the skeleton corresponding to the skeleton image according to the relevance judging result;
if the reliability judging result is that the second level is unreliable, determining whether the first preset number of classification results contain the first level or not;
if the first preset number of classification results comprise a first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
if the first classification results of the first preset number do not contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the second grade.
Optionally, the developmental grade determination module 304 may also be configured to:
if the correlation judgment result shows that the second level accords with the correlation rule, determining that the development level of the skeleton corresponding to the skeleton image is the second level;
if the correlation judgment result is that the second level does not accord with the correlation rule, determining whether the first level is contained in the first preset number of classification results;
if the first preset number of classification results comprise a first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
if the first classification results of the first preset number do not contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the second grade.
Optionally, the construction module 301 may be further configured to:
constructing a basic convolutional neural network model;
and constructing a first output branch and a second output branch at an output layer of the basic convolutional neural network model to obtain a skeleton development level detection model.
Optionally, in the training module 302, during training of the skeletal development level detection model, the loss function of the first output branch is cross entropy and the loss function of the second output branch is mean square error.
It will be clearly understood by those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in practical application, the above-mentioned functional allocation may be performed by different functional units and modules according to needs, i.e. the internal structure of the skeletal development level detection system is divided into different functional units or modules, so as to perform all or part of the above-mentioned functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above device may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
Fig. 4 is a schematic block diagram of a terminal device according to an embodiment of the present invention. As shown in fig. 4, the terminal device 40 of this embodiment includes: one or more processors 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. The processor 401, when executing the computer program 403, implements the steps of the above-described embodiments of the bone development level detection method, such as steps S101 to S104 shown in fig. 1. Alternatively, the processor 401 may implement the functions of the modules/units of the embodiment of the skeletal development level detection system described above, such as the functions of the modules 301 to 304 of fig. 3, when executing the computer program 403.
Illustratively, the computer program 403 may be partitioned into one or more modules/units that are stored in the memory 402 and executed by the processor 401 to complete the present application. The one or more modules/units may be a series of computer program instruction segments capable of performing the specified functions describing the execution of the computer program 403 in the terminal device 40. For example, the computer program 403 may be divided into a building module, a training module, a classification module and a development level determination module, each module specifically functioning as follows:
the building module is used for building a skeleton development grade detection model, an output layer of the skeleton development grade detection model comprises a first output branch and a second output branch, the first output branch adopts a softmax logistic regression method to conduct grade classification, and the second output branch adopts a linear regression method to conduct grade classification;
the training module is used for training the skeletal development level detection model to obtain a trained skeletal development level detection model;
the classification module is used for acquiring bone images, inputting the bone images into the trained bone development grade detection model, and obtaining a first classification result output by the first output branch and a second classification result output by the second output branch;
and the development grade determining module is used for determining the development grade of the skeleton corresponding to the skeleton image according to the first classification result and the second classification result.
Other modules or units may be described with reference to the embodiment shown in fig. 3, and will not be described here again.
The terminal device 40 may be a computing device such as a desktop computer, a notebook computer, a palm computer, and a cloud server. The terminal device 40 includes, but is not limited to, a processor 401, a memory 402. It will be appreciated by those skilled in the art that fig. 4 is merely an example of the terminal device 40 and is not meant to be limiting of the terminal device 40, and may include more or fewer components than shown, or may combine certain components, or different components, e.g., the terminal device 40 may also include input devices, output devices, network access devices, buses, etc.
The processor 401 may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 402 may be an internal storage unit of the terminal device 40, such as a hard disk or a memory of the terminal device 40. The memory 402 may also be an external storage device of the terminal device 40, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the terminal device 40. Further, the memory 402 may also include both internal storage units and external storage devices of the terminal device 40. The memory 402 is used for storing the computer program 403 and other programs and data required by the terminal device 40. The memory 402 may also be used to temporarily store data that has been output or is to be output.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiments provided herein, it should be understood that the disclosed skeletal development level detection system and method may be implemented in other ways. For example, the above-described embodiments of skeletal development level detection systems are merely illustrative, e.g., the division of the modules or units is merely a logical functional division, and may be implemented in other ways, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present application may implement all or part of the flow of the method of the above embodiment, or may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium, and when the computer program is executed by a processor, the computer program may implement the steps of each method embodiment described above. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium may include content that is subject to appropriate increases and decreases as required by jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is not included as electrical carrier signals and telecommunication signals.
The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.

Claims (9)

1. A method for detecting skeletal development levels, comprising:
constructing a skeletal development grade detection model, wherein an output layer of the skeletal development grade detection model comprises a first output branch and a second output branch, the first output branch adopts a softmax logistic regression method to carry out grade classification, and the second output branch adopts a linear regression method to carry out grade classification;
training the skeletal development level detection model to obtain a trained skeletal development level detection model;
acquiring a bone image, inputting the bone image into the trained bone development grade detection model, and obtaining a first classification result output by the first output branch and a second classification result output by the second output branch;
determining the development grade of the skeleton corresponding to the skeleton image according to the first classification result and the second classification result;
the first classification results comprise a first preset number of classification results and confidence degrees corresponding to the first preset number of classification results, wherein the first preset number of classification results are the classification results of the first preset number in the sequence from the high confidence degree to the low confidence degree in all classification results output by the first output branch; the second classification result is a floating point number;
the determining the development grade of the bone corresponding to the bone image according to the first classification result and the second classification result comprises the following steps:
converting the floating point number into a corresponding development grade, and marking the development grade as a first grade;
selecting a classification result with highest confidence from the first preset number of classification results, and marking the classification result as a second grade;
if the first level is the same as the second level, determining that the development level of the skeleton corresponding to the skeleton image is the first level;
if the first level is different from the second level, judging whether the second level is reliable or not, obtaining a reliability judging result, and determining the development level of the skeleton corresponding to the skeleton image according to the reliability judging result.
2. The skeletal development level detection method of claim 1, wherein said converting the floating point number to a corresponding development level, noted as a first level, comprises:
sequentially converting each development grade of the skeleton into a natural number starting from 0 according to the sequence from small to large to obtain a natural number grade sequence;
and determining the closest target natural number of the floating point number from the natural number grade sequence, and marking the development grade corresponding to the target natural number as a first grade.
3. The bone development level detection method according to claim 1, wherein the determining whether the second level is reliable or not, to obtain a reliability determination result, includes:
determining whether the second level has a suffix;
if the second level has no suffix, executing a first operation based on the first preset number of classification results and the confidence level of the second level to obtain a reliability judgment result;
if the second level has a suffix, determining a target bone age standard;
if the target bone age standard is RUS-CHN bone age standard, executing a first operation based on the first preset number of classification results and the second level of confidence coefficient to obtain a reliability judging result;
and if the target bone age standard is the bone age standard without the suffix, executing a second operation based on the classification results of the first preset number to obtain a reliability judgment result.
4. The bone development level detection method according to claim 3, wherein the performing a first operation based on the first preset number of classification results and the second level of confidence to obtain a reliability determination result includes:
determining whether the second level of confidence is greater than a confidence threshold;
if the confidence coefficient of the second level is larger than the confidence coefficient threshold value, the reliability judgment result is that the second level is reliable;
if the confidence coefficient of the second level is not greater than the confidence coefficient threshold value, determining whether the levels of the first preset number of classification results are continuous;
if the first preset number of classification results are continuous in level, the reliability judgment result is a second level reliable;
if the grades of the classification results of the first preset number are discontinuous, the reliability judgment result is that the second grade is unreliable;
and executing a second operation based on the classification results of the first preset number to obtain a reliability judgment result, wherein the method comprises the following steps:
determining whether the grades of the first preset number of classification results are continuous;
if the first preset number of classification results are continuous in level, the reliability judgment result is a second level reliable;
if the grades of the classification results of the first preset number are discontinuous, the reliability judgment result is that the second grade is unreliable.
5. The bone development level detection method according to claim 1, wherein the determining the development level of the bone corresponding to the bone image according to the reliability determination result includes:
if the reliability judging result is that the second level is reliable, judging whether the second level accords with a relevance rule, obtaining a relevance judging result, and determining a development level of a skeleton corresponding to the skeleton image according to the relevance judging result;
if the reliability judging result is that the second level is unreliable, determining whether the first preset number of classification results contain the first level or not;
if the first classification results of the first preset number contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
and if the first classification results of the first preset number do not contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the second grade.
6. The bone development level detection method according to claim 5, wherein the determining the development level of the bone corresponding to the bone image according to the correlation determination result includes:
if the correlation judgment result shows that the second level accords with the correlation rule, determining that the development level of the skeleton corresponding to the skeleton image is the second level;
if the correlation judgment result is that the second level does not accord with the correlation rule, determining whether the first level is contained in the first preset number of classification results;
if the first classification results of the first preset number contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the first grade;
and if the first classification results of the first preset number do not contain the first grade, determining that the development grade of the skeleton corresponding to the skeleton image is the second grade.
7. The method of any one of claims 1 to 6, wherein constructing a skeletal development level detection model comprises:
constructing a basic convolutional neural network model;
and constructing the first output branch and the second output branch at the output layer of the basic convolutional neural network model to obtain a skeletal development level detection model.
8. The bone development level detection method according to any one of claims 1 to 6, wherein the loss function of the first output branch is cross entropy and the loss function of the second output branch is mean square error in training the bone development level detection model.
9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the bone development level detection method according to any one of claims 1 to 8 when the computer program is executed.
CN202011015997.XA 2020-09-24 2020-09-24 Skeletal development grade detection method and terminal equipment Active CN112102942B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011015997.XA CN112102942B (en) 2020-09-24 2020-09-24 Skeletal development grade detection method and terminal equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011015997.XA CN112102942B (en) 2020-09-24 2020-09-24 Skeletal development grade detection method and terminal equipment

Publications (2)

Publication Number Publication Date
CN112102942A CN112102942A (en) 2020-12-18
CN112102942B true CN112102942B (en) 2024-04-05

Family

ID=73756046

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011015997.XA Active CN112102942B (en) 2020-09-24 2020-09-24 Skeletal development grade detection method and terminal equipment

Country Status (1)

Country Link
CN (1) CN112102942B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114822837A (en) * 2022-03-14 2022-07-29 合肥工业大学 Hand bone age automatic evaluation and correction method based on TW3 bone development association rule

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107591200A (en) * 2017-08-25 2018-01-16 卫宁健康科技集团股份有限公司 Stone age marker recognition appraisal procedure and system based on deep learning and image group
CN110660484A (en) * 2019-08-01 2020-01-07 平安科技（深圳）有限公司 Bone age prediction method, device, medium, and electronic apparatus
CN111209962A (en) * 2020-01-06 2020-05-29 电子科技大学 Combined image classification method based on CNN (CNN) feature extraction network) and combined heat map feature regression

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10588589B2 (en) * 2014-07-21 2020-03-17 Zebra Medical Vision Ltd. Systems and methods for prediction of osteoporotic fracture risk

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107591200A (en) * 2017-08-25 2018-01-16 卫宁健康科技集团股份有限公司 Stone age marker recognition appraisal procedure and system based on deep learning and image group
CN110660484A (en) * 2019-08-01 2020-01-07 平安科技（深圳）有限公司 Bone age prediction method, device, medium, and electronic apparatus
CN111209962A (en) * 2020-01-06 2020-05-29 电子科技大学 Combined image classification method based on CNN (CNN) feature extraction network) and combined heat map feature regression

Also Published As

Publication number Publication date
CN112102942A (en) 2020-12-18

Similar Documents

Publication Publication Date Title
CN109886127B (en) Fingerprint identification method and terminal equipment
CN113032580B (en) Associated file recommendation method and system and electronic equipment
CN113435499B (en) Label classification method, device, electronic equipment and storage medium
CN113298152B (en) Model training method, device, terminal equipment and computer readable storage medium
CN112214402B (en) Code verification algorithm selection method, device and storage medium
CN112488297B (en) Neural network pruning method, model generation method and device
CN111651674B (en) Bidirectional searching method and device and electronic equipment
CN112131274A (en) Method, device and equipment for detecting time series abnormal points and readable storage medium
CN112102942B (en) Skeletal development grade detection method and terminal equipment
CN111159481B (en) Edge prediction method and device for graph data and terminal equipment
CN110472227B (en) Vocabulary mastering degree evaluation method and device, terminal equipment and storage medium
CN116049345A (en) Document-level event joint extraction method and system based on bidirectional event complete graph
CN117851883B (en) Cross-modal large language model-based scene text detection and recognition method
CN115100739A (en) Man-machine behavior detection method, system, terminal device and storage medium
CN113010785B (en) User recommendation method and device
CN111563504A (en) License plate recognition method and related equipment
CN116662904A (en) Method, device, computer equipment and medium for detecting variation of data type
CN116911824A (en) Intelligent decision method and system based on electric power big data
CN116432608A (en) Text generation method and device based on artificial intelligence, computer equipment and medium
CN116016692A (en) Protocol description text construction method, device, equipment and storage medium
CN113449062B (en) Track processing method, track processing device, electronic equipment and storage medium
CN112990466A (en) Redundancy rule detection method and device and server
CN113836297A (en) Training method and device for text emotion analysis model
CN117132177B (en) Runoff forecasting model construction and runoff forecasting method based on multiple hypothesis test
CN112364581B (en) Method and device for automatically inserting specific codes into register transmission level design file

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant