CN115618282A - Synthetic gem identification method, device and storage medium - Google Patents
Synthetic gem identification method, device and storage medium Download PDFInfo
- Publication number
- CN115618282A CN115618282A CN202211617383.8A CN202211617383A CN115618282A CN 115618282 A CN115618282 A CN 115618282A CN 202211617383 A CN202211617383 A CN 202211617383A CN 115618282 A CN115618282 A CN 115618282A
- Authority
- CN
- China
- Prior art keywords
- gemstone
- data
- layer
- characteristic data
- synthetic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 239000010437 gem Substances 0.000 claims abstract description 193
- 229910001751 gemstone Inorganic materials 0.000 claims abstract description 175
- 238000012706 support-vector machine Methods 0.000 claims abstract description 48
- 238000000605 extraction Methods 0.000 claims abstract description 27
- 239000010410 layer Substances 0.000 claims description 76
- 238000012549 training Methods 0.000 claims description 73
- 239000002356 single layer Substances 0.000 claims description 60
- 230000006870 function Effects 0.000 claims description 47
- 239000013598 vector Substances 0.000 claims description 29
- 238000004590 computer program Methods 0.000 claims description 16
- 238000012795 verification Methods 0.000 claims description 15
- 238000011478 gradient descent method Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 description 19
- 238000004458 analytical method Methods 0.000 description 9
- 238000004364 calculation method Methods 0.000 description 9
- 239000011009 synthetic ruby Substances 0.000 description 8
- 238000012545 processing Methods 0.000 description 7
- 239000010979 ruby Substances 0.000 description 7
- 229910001750 ruby Inorganic materials 0.000 description 7
- 239000000126 substance Substances 0.000 description 7
- 230000003595 spectral effect Effects 0.000 description 5
- 229910052729 chemical element Inorganic materials 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 238000002329 infrared spectrum Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 235000013619 trace mineral Nutrition 0.000 description 4
- 239000011573 trace mineral Substances 0.000 description 4
- 241000579895 Chlorostilbon Species 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 239000010976 emerald Substances 0.000 description 3
- 229910052876 emerald Inorganic materials 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 238000002371 ultraviolet--visible spectrum Methods 0.000 description 3
- 238000001237 Raman spectrum Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 2
- 230000003321 amplification Effects 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- BASFCYQUMIYNBI-UHFFFAOYSA-N platinum Chemical compound [Pt] BASFCYQUMIYNBI-UHFFFAOYSA-N 0.000 description 2
- 229910052594 sapphire Inorganic materials 0.000 description 2
- 239000010980 sapphire Substances 0.000 description 2
- 229910018072 Al 2 O 3 Inorganic materials 0.000 description 1
- 238000000563 Verneuil process Methods 0.000 description 1
- 239000002969 artificial stone Substances 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000009658 destructive testing Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 238000001027 hydrothermal synthesis Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 229910052697 platinum Inorganic materials 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000001308 synthesis method Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 238000010189 synthetic method Methods 0.000 description 1
Images
Classifications
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a synthetic gem identification method, a synthetic gem identification device and a synthetic gem storage medium, wherein the synthetic gem identification method comprises the following steps: acquiring original characteristic data of a target gemstone, and inputting the original characteristic data into a multilayer perceptron to enable the multilayer perceptron to perform characteristic extraction on the original characteristic data to obtain classified characteristic data; inputting the classified feature data into a support vector machine, so that the support vector machine calculates a classification decision function according to the classified feature data and a hyperplane equation to obtain an identification result of the target gemstone, wherein the identification result comprises: the target gemstone is a synthetic gemstone or a natural gemstone, and the original characteristic data of the target gemstone is analyzed and operated through the multilayer sensing machine and the support vector machine so as to realize the automatic identification of the synthetic gemstone.
Description
Technical Field
The invention relates to the technical field of gem identification, in particular to a synthetic gem identification method, a synthetic gem identification device and a storage medium.
Background
The current main detection means of the identification of the synthesized gem comprises microscopic amplification, conventional instrument parameters, infrared spectrum test, ultraviolet visible spectrum test and the like, and then comprehensive judgment is carried out by detection personnel to give a result. Due to the wide variety of synthetic methods, there are different methods of identification for different methods, such as: the identification of the sapphire synthesized by the fluxing agent is based on the fact that the internal inclusion is fluxing agent residue, platinum sheet, strong fluorescence and 460nm and 470nm of ultraviolet visible possible deletion; the identification basis for synthesizing the emerald by a hydrothermal method comprises saw-toothed growth lines, a nail-shaped bag body and an infrared spectrum which have certain difference with the natural emerald; these criteria do not necessarily exist, however, and the test person may make a comprehensive empirical determination based on one or both of the unobvious identifying characteristics. Overall, the whole identification process is lack of systematicness, and the detection process is complex, tedious, time-consuming and labor-consuming; and certain subjective experience judgment components are provided, and misjudgment may occur in the identification result. Meanwhile, with the increasing development of synthetic technology, the identification of synthetic gems is more and more difficult, gems with unobvious identification characteristics often appear, and the identification becomes difficult without the need of next step or identification.
Disclosure of Invention
The invention provides a synthetic gem identification method, a synthetic gem identification device and a storage medium, wherein the synthetic gem automatic identification method is realized by analyzing and operating the original characteristic data of a target gem through a multilayer perceptron and a support vector machine.
In order to realize the automatic identification of the synthetic gemstones, the embodiment of the invention provides an identification method of the synthetic gemstones, which comprises the following steps: acquiring original characteristic data of a target gemstone, and inputting the original characteristic data into a multilayer perceptron to enable the multilayer perceptron to perform characteristic extraction on the original characteristic data to obtain classified characteristic data;
inputting the classified feature data into a support vector machine, so that the support vector machine calculates a classification decision function according to the classified feature data and a hyperplane equation to obtain an identification result of the target gemstone, wherein the identification result comprises: the target gemstone is a synthetic gemstone or the target gemstone is a natural gemstone.
As a preferred scheme, the synthetic gemstone identification method of the invention utilizes a multilayer perceptron to extract and screen the original characteristic data of a target gemstone, the classification characteristic data extracted by the multilayer perceptron has higher classification accuracy than the original characteristic data, then the classification characteristic data is input into a support vector machine to identify whether the target gemstone is a synthetic gemstone or a natural gemstone, a user tests the original characteristic of the target gemstone to obtain the original characteristic data, and the original characteristic data is input into an identification system, so that the identification system performs analysis and classification calculation processing on the original characteristic data through the multilayer perceptron and the support vector machine to obtain an identification result.
As a preferred scheme, the original feature data is input into a multilayer perceptron, so that the multilayer perceptron performs feature extraction on the original feature data to obtain classification feature data, specifically:
setting weight for original characteristic data; inputting the original feature data into a multilayer perceptron according to the weight;
the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, the original characteristic data is used as the input of a first layer of single-layer full-connection module, and the output of an upper layer of single-layer full-connection module is used as the input of a lower layer of single-layer full-connection module; outputting a preliminary identification result of the target gem by the last layer of single-layer full-connection module;
and acquiring the output of the first preset layer single-layer full-connection module as classification characteristic data.
As a preferred scheme, the multilayer sensing machine of the synthetic gemstone identification method extracts the features by using the plurality of single-layer full-connection modules, the extracted classification feature data has higher classification accuracy than the original feature data through the analysis and calculation of the multilayer single-layer full-connection modules on the original feature data, the identification result obtained by inputting the classification feature data into the support vector machine is more reliable, and the identification accuracy of the synthetic gemstone is improved.
As a preferred scheme, the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, which specifically comprise:
the first four layers of single-layer full-connection modules of the multilayer perceptron are as follows:
the last layer of single-layer full-connection module of the multilayer perceptron is as follows:
wherein ,
the sizes of the hidden layers from 1 st to nth layers;
the ith original characteristic data;
the ith input of the nth layer of single-layer full-connection module;
the ith weight vector is the nth single-layer full-connection module; n is a natural number;
as a preferred scheme, the multilayer sensing machine of the synthetic gemstone identification method extracts the features by using the plurality of single-layer full-connection modules, the extracted classification feature data has higher classification accuracy than the original feature data through the analysis and calculation of the multilayer single-layer full-connection modules on the original feature data, the identification result obtained by inputting the classification feature data into the support vector machine is more reliable, and the identification accuracy of the synthetic gemstone is improved.
Preferably, before inputting the raw feature data into the multi-layer perceptron, the method further includes:
acquiring original characteristic data of a plurality of natural gems and synthetic gems as training data, and setting corresponding verification data;
setting a plurality of weights for training data; inputting the training data into an initial multi-layer perceptron according to the weight;
calculating partial derivatives of the loss functions to the weight parameters of the initial multilayer perceptron by utilizing back propagation to obtain the gradient of the weight parameters, and updating the weight parameters by using a random gradient descent method; and until the initial multilayer perceptron calculates and verifies that the accuracy is higher than a preset value according to the verification data, and the trained multilayer perceptron is obtained.
As a preferred scheme, before the original characteristic data is input into the multilayer perceptron, the original multilayer perceptron is trained by using the original characteristic data of a plurality of natural gems and synthetic gems, the weight parameters of the multilayer perceptron are optimized, and the classification accuracy of the multilayer perceptron is improved.
As a preferred scheme, the support vector machine calculates a classification decision function according to the classification feature data and the hyperplane equation to obtain the identification result of the target gemstone, and specifically comprises the following steps:
inputting the feature vectors of the classified feature data into a hyperplane equation of a support vector machine, wherein the hyperplane equation is calculated from a training data set;
the hyperplane equation is:
wherein ,
and
calculating from the training data set a hyperplane parameter;
and further calculating a classification decision function, wherein the classification decision function is as follows:
wherein ,
is a sign function;
obtaining an identification result of the target gem according to the result of the classification decision function;
when in use
The target gemstone is a synthetic gemstone;
when in use
The target gemstone is a natural gemstone.
As a preferred scheme, the method inputs the classification characteristic data into the support vector machine to identify whether the target gemstone is a synthetic gemstone or a natural gemstone, the user tests the original characteristic of the target gemstone to obtain the original characteristic data, and inputs the original characteristic data into the identification system, so that the identification system obtains an identification result by the calculation processing of the analysis and classification of the original characteristic data through the multilayer perceptron and the support vector machine.
As a preferred scheme, before inputting the classification feature data into a support vector machine, the method further comprises:
constructing a training data set as
；
wherein 
，
For the feature vector of the nth classified feature data,
the designation +1 is for a synthetic gemstone,
1 represents a natural gemstone;
according to penalty parameter
And calculating a quadratic programming problem:
wherein ,
and
is a lagrange multiplier;
is the feature vector of the ith classification feature data,
feature vector for jth classification feature data
Get the optimal solution
；
Computing hyperplane parameters from an optimal solution
And
；
selecting
A component of
Satisfies the conditions
Calculating
Obtaining a hyperplane equation:
as a preferred scheme, before the classification characteristic data is input into a support vector machine, a training data set is constructed to calculate parameters of a hyperplane, a hyperplane equation is obtained, a support vector machine model with the best classification effect is obtained, the classification characteristic data input into the target gemstone is processed, and finally the gemstone is identified according to a classification decision function to obtain an identification result.
Accordingly, the present invention also provides an apparatus for authenticating a synthetic gemstone, comprising: the device comprises a feature extraction module and an identification module;
the characteristic extraction module is used for acquiring original characteristic data of a target gemstone and inputting the original characteristic data into the multilayer perceptron so as to enable the multilayer perceptron to extract characteristics of the original characteristic data and acquire classified characteristic data;
the identification module is used for inputting the classification characteristic data into a support vector machine, so that the support vector machine calculates a classification decision function according to the classification characteristic data and a hyperplane equation to obtain an identification result of the target gemstone, wherein the identification result comprises: the target gemstone is a synthetic gemstone or the target gemstone is a natural gemstone.
As a preferred scheme, the characteristic extraction module of the synthetic gemstone identification device of the invention utilizes a multilayer perceptron to extract and screen the original characteristic data of a target gemstone, the classification characteristic data extracted by the multilayer perceptron has higher classification accuracy than the original characteristic data, the identification module inputs the classification characteristic data into a support vector machine to identify whether the target gemstone is a synthetic gemstone or a natural gemstone, a user tests the original characteristic of the target gemstone to obtain the original characteristic data, and inputs the original characteristic data into an identification system, so that the identification system performs analysis and classification calculation processing on the original characteristic data through the multilayer perceptron and the support vector machine to obtain an identification result.
Preferably, the feature extraction module includes: a first training unit and an extraction unit;
the first training unit is used for acquiring original characteristic data of a plurality of natural gems and synthetic gems as training data and setting corresponding verification data; setting a plurality of weight parameters for training data; inputting the training data into an initial multi-layer perceptron according to the weight parameters; calculating partial derivatives of the loss function to the weight parameters of the initial multilayer perceptron by utilizing back propagation to obtain the gradient of the weight parameters, and updating the weight parameters by using a random gradient descent method; until the initial multilayer perceptron calculates that the verification accuracy is higher than a preset value according to the verification data, and the trained multilayer perceptron is obtained;
the extraction unit is used for setting weight to the original characteristic data; inputting the original feature data into a multilayer perceptron according to the weight; the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, the original characteristic data is used as the input of a first layer of single-layer full-connection module, and the output of an upper layer of single-layer full-connection module is used as the input of a lower layer of single-layer full-connection module; outputting a preliminary identification result of the target gem by the last layer of single-layer full-connection module; and acquiring the output of the first preset layer single-layer full-connection module as classification characteristic data.
As a preferred scheme, the extraction module of the synthetic gemstone identification device extracts the features by utilizing a plurality of single-layer full-connection modules of the multilayer perceptron, the extracted classification feature data has higher classification accuracy than the original feature data through the analysis and calculation of the original feature data by the multilayer single-layer full-connection modules, the identification result obtained by inputting the classification feature data into the support vector machine is more reliable, and the identification accuracy of the synthetic gemstone is improved.
Preferably, the identification module comprises a second training unit and an identification unit;
the second training unit is used for constructing a training data set as
；
wherein 
，
For the feature vector of the nth classified feature data,
the designation +1 is for a synthetic gemstone,
1 represents a natural gemstone;
according to penalty parameter
Calculating a quadratic programming problem:
wherein ,
and
is a lagrange multiplier;
is the feature vector of the ith classification feature data,
feature vector for jth classification feature data
Obtaining an optimal solution
；
Computing hyperplane parameters from an optimal solution
And
；
selecting
A component of
Satisfy the condition
Calculating
Obtaining a hyperplane equation:
the identification unit is used for inputting the feature vectors of the classified feature data into a hyperplane equation of a support vector machine, and the hyperplane equation is calculated from a training data set;
the hyperplane equation is:
wherein ,
and
calculating from the training data set a hyperplane parameter;
and further calculating a classification decision function, wherein the classification decision function is as follows:
wherein ,
is a sign function;
obtaining an identification result of the target gem according to the result of the classification decision function;
when the temperature is higher than the set temperature
The target gemstone is a synthetic gemstone;
when in use
And the target gemstone is a natural gemstone.
As a preferred scheme, the identification module inputs the classified characteristic data into the support vector machine to identify whether the target gemstone is a synthetic gemstone or a natural gemstone, the user tests the original characteristic of the target gemstone to obtain the original characteristic data, and inputs the original characteristic data into the identification system, so that the identification system performs calculation processing of analysis and classification on the original characteristic data through the multilayer perceptron and the support vector machine to obtain an identification result.
Accordingly, the present invention also provides a computer readable storage medium comprising a stored computer program; wherein the computer program, when executed, controls an apparatus in which the computer readable storage medium is located to perform a method of synthetic gemstone authentication according to the present disclosure.
Drawings
FIG. 1 is a schematic flow chart diagram illustrating one embodiment of a method for identifying a synthetic gemstone according to the present invention;
fig. 2 is a schematic structural view of an embodiment of an apparatus for authenticating a synthetic gemstone according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
Referring to fig. 1, a method for identifying a synthetic gemstone according to an embodiment of the present invention includes steps S101-S102:
in this embodiment, the synthetic gemstones include synthetic ruby, synthetic sapphire, and synthetic emerald.
Step S101: the method comprises the steps of obtaining original characteristic data of a target gem, inputting the original characteristic data into a multilayer perceptron, and enabling the multilayer perceptron to conduct characteristic extraction on the original characteristic data to obtain classified characteristic data.
In this embodiment, the original feature data of the target gemstone specifically includes: carrying out amplification inspection, fluorescence test, infrared spectrum test, ultraviolet-visible spectrum test and Raman spectrum test on the target gemstone; and chemical composition testing; and obtaining data of characteristic absorption peaks, main chemical component content, trace elements and contents of the spectral data as original characteristic data.
In this embodiment, the original feature data is input into a multi-layer perceptron, so that the multi-layer perceptron performs feature extraction on the original feature data to obtain classification feature data, which specifically includes:
setting weight for the original characteristic data; inputting the original feature data into a multilayer perceptron according to the weight;
the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, the original characteristic data is used as the input of a first layer of single-layer full-connection module, and the output of an upper layer of single-layer full-connection module is used as the input of a lower layer of single-layer full-connection module; outputting a preliminary identification result of the target gem by the last layer of single-layer full-connection module;
and acquiring the output of the first preset layer single-layer full-connection module as classification characteristic data.
In this embodiment, the multilayer perceptron includes a plurality of layers of single-layer full-connection modules, specifically:
the first four layers of single-layer full-connection modules of the multilayer perceptron are as follows:
the last layer of single-layer full-connection module of the multilayer perceptron is as follows:
wherein ,
the sizes of the hidden layers from 1 st to nth layers;
the ith original characteristic data;
the ith input of the nth layer single-layer full-connection module;
the ith weight vector is the nth single-layer full-connection module; n is a natural number;
in the present embodiment, the multilayer perceptron includes 5 layers of single-layer fully-connected modules, i.e. n =5;
the sizes of the different hidden layers are respectively:
obtaining output x of layer 4 single-layer full-connection module 4 As classification characteristic data.
In this embodiment, before inputting the raw feature data into the multi-layer perceptron, the method further includes:
acquiring original characteristic data of a plurality of natural gems and synthetic gems as training data, and setting corresponding verification data;
setting a plurality of weights for training data; inputting the training data into an initial multi-layer perceptron according to the weight;
calculating partial derivatives of the loss functions to the weight parameters of the initial multilayer perceptron by utilizing back propagation to obtain the gradient of the weight parameters, and updating the weight parameters by using a random gradient descent method; and until the initial multi-layer perceptron calculates and verifies that the accuracy is higher than a preset value according to the verification data, and the trained multi-layer perceptron is obtained.
In this example, natural gems of 6 different places (Burma, vietnam, thailand, mosangbisk, tanzania, mogarska) and synthetic gems synthesized by 4 synthesis methods (flame fusion, czochralski, hydrothermal, flux) were collected as training library samples;
and testing the training library samples to obtain inclusion data and spectral data of different types of samples, wherein the spectral data comprise fluorescence characteristic spectra, reflection and transmission characteristic infrared spectra, raman spectra and ultraviolet visible spectra, and reading characteristic absorption peaks and inclusion data of the spectral data as basic characteristic training data.
Illustratively, natural ruby and synthetic ruby of a training library sample are tested, and basic feature training data for obtaining spectral data of the natural ruby and synthetic ruby are as follows:
and (3) carrying out chemical component test on partial training library samples to obtain the main chemical components and contents and trace elements and contents of different types of samples as chemical characteristic training data.
Illustratively, destructive testing of a portion of natural ruby and a portion of synthetic ruby was conducted to determine that the major chemical constituent of natural ruby and synthetic ruby was Al 2 O 3 The contents of K, ca, cr, fe, ti, V, ga, zr, mo and Pb, and the main quantity of Al on the energy spectrum 2 O 3 And a nondestructive test curve of the trace elements; and (3) carrying out primary chemical component content, trace element content and content test on the natural ruby and the synthetic ruby of the training library sample by using energy spectrum test to obtain chemical element characteristic training data of the natural ruby and the synthetic ruby.
And taking the basic characteristic training data and the chemical element characteristic training data of the training library samples as the original characteristic data of the training library samples.
In this embodiment, the original feature data of the training library sample is obtained and stored as a numpy file, and a corresponding label document is constructed, where the label document of the synthetic gemstone is marked as 1, and the label document of the natural gemstone is marked as 0;
the basic feature training data and the chemical element feature training data are calculated by a ratio of 15%: inputting 85% of weight proportion into an initial multilayer perceptron, and obtaining an output value through forward propagation;
according to BP objective function
Obtaining the gradient of each weight parameter of the initial multilayer perceptron, and updating the weight parameter of the initial multilayer perceptron by using a random gradient descent method; until the initial multilayer perceptron calculates and verifies the accuracy rate to be higher than 80% according to the label file, obtain the multilayer perceptron after training;
where t is the desired output and z is the actual output. The desired output of the synthetic stone t =1 and the desired output of the natural stone t =0 are set.
The architecture of the initial multi-tier perceptron is as follows:
wherein, the sizes of different hidden layers are respectively as follows:
step S102: inputting the classified feature data into a support vector machine, so that the support vector machine calculates a classification decision function according to the classified feature data and a hyperplane equation to obtain an identification result of the target gemstone, wherein the identification result comprises: the target gemstone is a synthetic gemstone or the target gemstone is a natural gemstone.
In this embodiment, the support vector machine calculates a classification decision function according to the classification feature data and the hyperplane equation, and obtains an identification result of the target gemstone, specifically:
inputting the feature vectors of the classified feature data into a hyperplane equation of a support vector machine, wherein the hyperplane equation is calculated from a training data set;
the hyperplane equation is:
wherein ,
and
calculating from the training data set a hyperplane parameter;
and further calculating a classification decision function, wherein the classification decision function is as follows:
wherein ,
is a sign function;
obtaining an identification result of the target gemstone according to the result of the classification decision function;
when in use
The target gemstone is a synthetic gemstone;
when the temperature is higher than the set temperature
The target gemstone is a natural gemstone.
In this embodiment, the outputs of the support vector machine "+1" and "-1" are mapped to "synthetic ruby" and "natural ruby", respectively; and writing the identification result into a fixed text file.
In this embodiment, before inputting the classification feature data into the support vector machine, the method further includes:
constructing a training data set as
；
wherein 
，
For the feature vector of the nth classified feature data,
is given as a +1 representation of a synthetic gemstone,
is-1 represents a natural gemstone;
according to penalty parameter
Calculating a quadratic programming problem:
wherein ,
and
is a lagrange multiplier;
is the feature vector of the ith classification feature data,
feature vector for jth classification feature data
Obtaining an optimal solution
；
Calculating hyperplane parameters from the optimal solution
And
；
selecting
A component of
Satisfies the conditions
Calculating
Obtaining a hyperplane equation:
in the present embodiment, the basic feature training data and the chemical element feature training data are set at 15%: after 85% of weight proportion is input into the trained multilayer perceptron, the output of a plurality of 4 th layer single-layer full-connection modules is obtained and used as the training data of a training data set, and the training data set is constructed as
；
wherein ,
,
for the feature vector of the nth classified feature data,
the designation +1 is for a synthetic gemstone,
1 represents a natural gemstone;
setting penalty parameter set
Use of
Respectively solving a quadratic programming problem, training the support vector machine, selecting a hyperplane equation corresponding to the optimal solution as a hyperplane square of the trained support vector machineThe process.
In this embodiment, the training targets of the support vector machine are: for a given data set T and hyperplane
Maximizing the geometrical spacing of the hyperplane with respect to all sample points
；
wherein ,
can be simplified to a convex quadratic programming problem with inequality constraints, i.e.
Using a lagrange multiplier method, constructing a lagrange objective function as:
wherein ,
in order to be a lagrange multiplier,
。
the final optimization objective was:
the embodiment of the invention has the following effects:
the synthetic gemstone identification method of the invention utilizes the multilayer perceptron to extract and screen the original characteristic data of the target gemstone, the classification characteristic data extracted by the multilayer perceptron has higher classification accuracy than the original characteristic data, then the classification characteristic data is input into the support vector machine to identify whether the target gemstone is a synthetic gemstone or a natural gemstone, a user tests the original characteristic of the target gemstone to obtain the original characteristic data, and the original characteristic data is input into the identification system, so that the identification system performs the analysis and classification calculation processing on the original characteristic data through the multilayer perceptron and the support vector machine to obtain the identification result.
Example two
Referring to fig. 2, an apparatus for identifying a synthetic gemstone according to an embodiment of the present invention includes: a feature extraction module 201 and an identification module 202;
the feature extraction module 201 is configured to obtain original feature data of a target gemstone, and input the original feature data into a multilayer perceiving machine, so that the multilayer perceiving machine performs feature extraction on the original feature data to obtain classified feature data;
the identification module 202 is configured to input the classification feature data into a support vector machine, so that the support vector machine calculates a classification decision function according to the classification feature data and a hyperplane equation, and obtains an identification result of the target gemstone, where the identification result includes: the target gemstone is a synthetic gemstone or the target gemstone is a natural gemstone.
The feature extraction module includes: a first training unit and an extraction unit;
the first training unit is used for acquiring original characteristic data of a plurality of natural gems and synthetic gems as training data and setting corresponding verification data; setting a plurality of weight parameters for training data; inputting the training data into an initial multi-layer perceptron according to the weight parameters; calculating partial derivatives of the loss functions to the weight parameters of the initial multilayer perceptron by utilizing back propagation to obtain the gradient of the weight parameters, and updating the weight parameters by using a random gradient descent method; until the initial multilayer perceptron calculates that the verification accuracy is higher than a preset value according to the verification data, and the trained multilayer perceptron is obtained;
the extraction unit is used for setting weight for the original characteristic data; inputting the original feature data into a multilayer perceptron according to the weight; the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, the original characteristic data is used as the input of a first layer of single-layer full-connection module, and the output of an upper layer of single-layer full-connection module is used as the input of a lower layer of single-layer full-connection module; outputting a preliminary identification result of the target gem by the last layer of single-layer full-connection module; and acquiring the output of the first preset layer single-layer full-connection module as classification characteristic data.
The identification module comprises a second training unit and an identification unit;
the second training unit is used for constructing a training data set as
；
wherein 
，
For the feature vector of the nth classified feature data,
the designation +1 is for a synthetic gemstone,
1 represents a natural gemstone;
according to penalty parameter
Calculating a quadratic programming problem:
wherein ,
and
is a lagrange multiplier;
is the feature vector of the ith classification feature data,
feature vector for jth classification feature data
Obtaining an optimal solution
；
Calculating hyperplane parameters from the optimal solution
And
；
selecting
A component of
Satisfies the conditions
Calculating
Obtaining a hyperplane equation:
the identification unit is used for inputting the feature vectors of the classified feature data into a hyperplane equation of a support vector machine, and the hyperplane equation is calculated from a training data set;
the hyperplane equation is:
wherein ,
and
calculated from the training data set as hyperplane parameters；
And further calculating a classification decision function, wherein the classification decision function is as follows:
wherein ,
is a sign function;
obtaining an identification result of the target gem according to the result of the classification decision function;
when in use
The target gemstone is a synthetic gemstone;
when the temperature is higher than the set temperature
The target gemstone is a natural gemstone.
The synthetic gemstone authentication device may implement the synthetic gemstone authentication method of the above method embodiments. The alternatives in the above-described method embodiments are also applicable to this embodiment and will not be described in detail here. The rest of the embodiments of the present application may refer to the contents of the above method embodiments, and in this embodiment, details are not repeated.
The embodiment of the invention has the following effects:
the characteristic extraction module of the synthetic gemstone identification device extracts and screens original characteristic data of a target gemstone by using the multilayer perceptron, the classification characteristic data extracted by the multilayer perceptron has higher classification accuracy than the original characteristic data, the identification module inputs the classification characteristic data into the support vector machine to identify whether the target gemstone is a synthetic gemstone or a natural gemstone, and a user tests the original characteristic of the target gemstone to obtain the original characteristic data and inputs the original characteristic data into the identification system, so that the identification system performs analysis and classification calculation processing on the original characteristic data by using the multilayer perceptron and the support vector machine to obtain an identification result.
EXAMPLE III
Accordingly, the present invention also provides a computer readable storage medium comprising a stored computer program, wherein the computer program when executed controls an apparatus in which the computer readable storage medium is located to perform a method of authenticating a synthetic gemstone according to any one of the embodiments above.
Illustratively, the computer program may be partitioned into one or more modules/units that are stored in the memory and executed by the processor to implement the invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used for describing the execution process of the computer program in the terminal device.
The terminal device can be a desktop computer, a notebook, a palm computer, a cloud server and other computing devices. The terminal device may include, but is not limited to, a processor, a memory.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, or the like. The general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like, said processor being the control center of said terminal device, and various interfaces and lines are used to connect the various parts of the whole terminal device.
The memory may be used to store the computer programs and/or modules, and the processor may implement various functions of the terminal device by executing or executing the computer programs and/or modules stored in the memory and calling data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the mobile terminal, and the like. In addition, the memory may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
Wherein, the terminal device integrated module/unit can be stored in a computer readable storage medium if it is implemented in the form of software functional unit and sold or used as a stand-alone product. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments described above may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like.
The above-mentioned embodiments are provided to further explain the objects, technical solutions and advantages of the present invention in detail, and it should be understood that the above-mentioned embodiments are only examples of the present invention and are not intended to limit the scope of the present invention. It should be understood that any modifications, equivalents, improvements and the like, which come within the spirit and principle of the invention, may occur to those skilled in the art and are intended to be included within the scope of the invention.
Claims (10)
1. A method of identifying a synthetic gemstone, comprising:
acquiring original characteristic data of a target gemstone, and inputting the original characteristic data into a multilayer perceptron to enable the multilayer perceptron to perform characteristic extraction on the original characteristic data to obtain classified characteristic data;
inputting the classified feature data into a support vector machine, so that the support vector machine calculates a classification decision function according to the classified feature data and a hyperplane equation to obtain an identification result of the target gemstone, wherein the identification result comprises: the target gemstone is a synthetic gemstone or the target gemstone is a natural gemstone.
2. A method for authenticating a synthetic gemstone according to claim 1, wherein said raw characteristic data is inputted into a multi-layered perceiver, so that said multi-layered perceiver performs characteristic extraction on said raw characteristic data to obtain classified characteristic data, in particular:
setting weight for the original characteristic data; inputting the original feature data into a multilayer perceptron according to the weight;
the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, the original characteristic data is used as the input of a first layer of single-layer full-connection module, and the output of an upper layer of single-layer full-connection module is used as the input of a lower layer of single-layer full-connection module; outputting a preliminary identification result of the target gem by the last layer of single-layer full-connection module;
and acquiring the output of the first preset layer single-layer full-connection module as classification characteristic data.
3. A method of authenticating a synthetic gemstone according to claim 2, wherein said multilayer perceptron comprises a plurality of layers of single-layer fully-connected modules, in particular:
the first four layers of single-layer full-connection modules of the multilayer perceptron are as follows:
the last layer of single-layer full-connection module of the multilayer perceptron is as follows:
wherein ,
the sizes of the hidden layers of the 1 st layer to the nth layer;
the ith original characteristic data;
the ith input of the nth layer single-layer full-connection module;
the ith weight vector is the nth single-layer full-connection module; n is a natural number;
4. a method of authenticating a synthetic gemstone according to claim 1, wherein said inputting said raw characteristic data to a multilayer perceptron further comprises:
acquiring original characteristic data of a plurality of natural gems and synthetic gems as training data, and setting corresponding verification data;
setting a plurality of weights for training data; inputting the training data into an initial multi-layer perceptron according to the weights;
calculating partial derivatives of the loss functions to the weight parameters of the initial multilayer perceptron by utilizing back propagation to obtain the gradient of the weight parameters, and updating the weight parameters by using a random gradient descent method; and until the initial multi-layer perceptron calculates and verifies that the accuracy is higher than a preset value according to the verification data, and the trained multi-layer perceptron is obtained.
5. The method of claim 1, wherein said support vector machine calculates a classification decision function based on said classification characteristic data and a hyperplane equation to obtain an identification result of the target gemstone, and specifically comprises:
inputting the feature vectors of the classified feature data into a hyperplane equation of a support vector machine, wherein the hyperplane equation is calculated from a training data set;
the hyperplane equation is:
wherein ,
and
calculating from the training data set a hyperplane parameter;
and further calculating a classification decision function, wherein the classification decision function is as follows:
wherein ,
is a sign function;
obtaining an identification result of the target gem according to the result of the classification decision function;
when in use
The target gemstone is a synthetic gemstone;
when in use
And the target gemstone is a natural gemstone.
6. A method of identifying a synthetic gemstone according to claim 1, wherein prior to inputting said classification characteristic data into a support vector machine, further comprising:
constructing a training data set as
；
wherein 
，
For the feature vector of the nth classified feature data,
the designation +1 is for a synthetic gemstone,
1 represents a natural gemstone;
according to penalty parameter
Calculating a quadratic programming problem:
wherein ,
and
is a lagrange multiplier;
is the feature vector of the ith classification feature data,
feature vector for jth classification feature data
Get the optimal solution
；
Computing hyperplane parameters from an optimal solution
And
；
selecting
A component of
Satisfies the conditions
Calculating
Obtaining a hyperplane equation:
7. an apparatus for authenticating a synthetic gemstone, comprising: the device comprises a feature extraction module and an identification module;
the characteristic extraction module is used for acquiring original characteristic data of a target gemstone and inputting the original characteristic data into the multilayer perceptron so as to enable the multilayer perceptron to perform characteristic extraction on the original characteristic data and acquire classified characteristic data;
the identification module is used for inputting the classification characteristic data into a support vector machine so that the support vector machine calculates a classification decision function according to the classification characteristic data and a hyperplane equation to obtain an identification result of the target gemstone, wherein the identification result comprises: the target gemstone is a synthetic gemstone or the target gemstone is a natural gemstone.
8. A synthetic gemstone authentication apparatus as recited in claim 7, wherein said feature extraction module includes: a first training unit and an extraction unit;
the first training unit is used for acquiring original characteristic data of a plurality of natural gems and synthetic gems as training data and setting corresponding verification data; setting a plurality of weight parameters for training data; inputting the training data into an initial multi-layer perceptron according to the weight parameters; calculating partial derivatives of the loss function to the weight parameters of the initial multilayer perceptron by utilizing back propagation to obtain the gradient of the weight parameters, and updating the weight parameters by using a random gradient descent method; until the initial multi-layer perceptron calculates the verification accuracy rate to be higher than the preset value according to the verification data, and the trained multi-layer perceptron is obtained;
the extraction unit is used for setting weight to the original characteristic data; inputting the original feature data into a multilayer perceptron according to the weight; the multilayer perceptron comprises a plurality of layers of single-layer full-connection modules, the original characteristic data is used as the input of a first layer of single-layer full-connection module, and the output of an upper layer of single-layer full-connection module is used as the input of a lower layer of single-layer full-connection module; outputting a preliminary identification result of the target gem by the last layer of single-layer full-connection module; and acquiring the output of the first preset layer single-layer full-connection module as classification characteristic data.
9. A synthetic gemstone authentication apparatus as recited in claim 7, wherein said authentication module includes a second training unit and an authentication unit;
the second training unit is used for constructing a training data set as
；
wherein 
，
For the feature vector of the nth classified feature data,
the designation +1 is for a synthetic gemstone,
1 represents a natural gemstone;
according to penalty parameter
And calculating a quadratic programming problem:
wherein ,
and
is a lagrange multiplier;
for the feature vector of the ith classification feature data,
feature vector for jth classification feature data
Obtaining an optimal solution
；
Computing hyperplane parameters from an optimal solution
And
；
selecting
A component of
Satisfies the conditions
Calculating
Obtaining a hyperplane equation:
the identification unit is used for inputting the feature vectors of the classified feature data into a hyperplane equation of a support vector machine, and the hyperplane equation is calculated from a training data set;
the hyperplane equation is:
wherein ,
and
calculating from the training data set a hyperplane parameter;
and further calculating a classification decision function, wherein the classification decision function is as follows:
wherein ,
is a sign function;
obtaining an identification result of the target gemstone according to the result of the classification decision function;
when in use
The target gemstone is a synthetic gemstone;
when the temperature is higher than the set temperature
And the target gemstone is a natural gemstone.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored computer program; wherein the computer program, when executed, controls an apparatus in which the computer readable storage medium is located to perform a method of authenticating a synthetic gemstone according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211617383.8A CN115618282B (en) | 2022-12-16 | 2022-12-16 | Identification method, device and storage medium for synthetic precious stone |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211617383.8A CN115618282B (en) | 2022-12-16 | 2022-12-16 | Identification method, device and storage medium for synthetic precious stone |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115618282A true CN115618282A (en) | 2023-01-17 |
| CN115618282B CN115618282B (en) | 2023-06-06 |
Family
ID=84881008
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211617383.8A Active CN115618282B (en) | 2022-12-16 | 2022-12-16 | Identification method, device and storage medium for synthetic precious stone |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115618282B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116777915A (en) * | 2023-08-23 | 2023-09-19 | 国检中心深圳珠宝检验实验室有限公司 | Method and device for identifying sub-class of sapphire and storage medium |
| CN116818761A (en) * | 2023-08-16 | 2023-09-29 | 国检中心深圳珠宝检验实验室有限公司 | Method and equipment for detecting color cause of single crystal color precious stone |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020221278A1 (en) * | 2019-04-29 | 2020-11-05 | 北京金山云网络技术有限公司 | Video classification method and model training method and apparatus thereof, and electronic device |
| CN114997287A (en) * | 2022-05-19 | 2022-09-02 | 浙江商汤科技开发有限公司 | Model training and data processing method, device, equipment and storage medium |
| CN115186776A (en) * | 2022-09-13 | 2022-10-14 | 国检中心深圳珠宝检验实验室有限公司 | Method, device and storage medium for classifying ruby producing areas |
| CN115205255A (en) * | 2022-07-15 | 2022-10-18 | 广东恒利混凝土制品有限公司 | Stone automatic grading method and system based on deep learning |
-
2022
- 2022-12-16 CN CN202211617383.8A patent/CN115618282B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020221278A1 (en) * | 2019-04-29 | 2020-11-05 | 北京金山云网络技术有限公司 | Video classification method and model training method and apparatus thereof, and electronic device |
| CN114997287A (en) * | 2022-05-19 | 2022-09-02 | 浙江商汤科技开发有限公司 | Model training and data processing method, device, equipment and storage medium |
| CN115205255A (en) * | 2022-07-15 | 2022-10-18 | 广东恒利混凝土制品有限公司 | Stone automatic grading method and system based on deep learning |
| CN115186776A (en) * | 2022-09-13 | 2022-10-14 | 国检中心深圳珠宝检验实验室有限公司 | Method, device and storage medium for classifying ruby producing areas |
Non-Patent Citations (1)
| Title |
|---|
| 乔淑君: "地质样品的激光诱导击穿光谱数据库和分类方法的研究" * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116818761A (en) * | 2023-08-16 | 2023-09-29 | 国检中心深圳珠宝检验实验室有限公司 | Method and equipment for detecting color cause of single crystal color precious stone |
| CN116818761B (en) * | 2023-08-16 | 2024-04-19 | 国检中心深圳珠宝检验实验室有限公司 | Method and equipment for detecting color cause of single crystal color precious stone |
| CN116777915A (en) * | 2023-08-23 | 2023-09-19 | 国检中心深圳珠宝检验实验室有限公司 | Method and device for identifying sub-class of sapphire and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115618282B (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115618282A (en) | Synthetic gem identification method, device and storage medium | |
| CN107690671B (en) | Method and system for the classification of iteration defect | |
| Feilhauer et al. | Multi-method ensemble selection of spectral bands related to leaf biochemistry | |
| CN103597346B (en) | Use the analogy method of Non-Destructive Testing operation under the full-scale condition of composite signal | |
| CN107820620A (en) | Method and system for defect classification | |
| Koch et al. | Automated drill core mineralogical characterization method for texture classification and modal mineralogy estimation for geometallurgy | |
| CN108596246A (en) | The method for building up of soil heavy metal content detection model based on deep neural network | |
| CN105043998B (en) | One kind differentiates the haploid method of corn | |
| US8532256B1 (en) | Method and apparatus for the automated assay and valuation of precious metal objects | |
| McMillan et al. | Correlation of limestone beds using laser-induced breakdown spectroscopy and chemometric analysis | |
| Jayakody et al. | A generalised approach for high-throughput instance segmentation of stomata in microscope images | |
| CN112215696A (en) | Personal credit evaluation and interpretation method, device, equipment and storage medium based on time sequence attribution analysis | |
| JP7076463B2 (en) | Spectrum analyzer and spectrum analysis method | |
| Marschalek et al. | Image-Rice Grain Scanner: a three-dimensional fully automated assessment of grain size and quality traits | |
| CN113008865A (en) | Method, device, medium and equipment for identifying jewelry jade | |
| Wang et al. | Corn seed defect detection based on watershed algorithm and two-pathway convolutional neural networks | |
| Chang et al. | Time-series growth prediction model based on U-net and machine learning in Arabidopsis | |
| CN105223140A (en) | The method for quickly identifying of homology material | |
| Duncan | Multi-Sensor Fusion Technology, Visual Imagery and Predictive Modeling Tools, and Big Geospatial Data Analytics in the Virtual Economy of the Metaverse | |
| CN115186776B (en) | Method, device and storage medium for classifying ruby producing areas | |
| Daunys et al. | Comparison of computer vision models in application to pollen classification using light scattering | |
| TW201321739A (en) | Signal analysis device, signal analysis method and computer program product | |
| De Castro et al. | Novel technique for the preparation and analysis of powder-based polished sections by automated optical mineralogy: Part 2–Use of deep learning approach for transparent mineral detection | |
| CN113125374B (en) | Method, device and equipment for detecting REE content in carbonate type rare earth deposit sample | |
| Melcher et al. | Analytical proof of origin for raw materials |
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 |