CN114998639A - Chinese medicinal material class identification method based on deep learning - Google Patents
Chinese medicinal material class identification method based on deep learning Download PDFInfo
- Publication number
- CN114998639A CN114998639A CN202210412607.5A CN202210412607A CN114998639A CN 114998639 A CN114998639 A CN 114998639A CN 202210412607 A CN202210412607 A CN 202210412607A CN 114998639 A CN114998639 A CN 114998639A
- Authority
- CN
- China
- Prior art keywords
- traditional chinese
- herbs
- chinese medicine
- image
- chinese medicinal
- 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
- 239000000463 material Substances 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000013135 deep learning Methods 0.000 title claims abstract description 21
- 239000003814 drug Substances 0.000 claims abstract description 74
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000001914 filtration Methods 0.000 claims abstract description 8
- 230000004927 fusion Effects 0.000 claims abstract description 8
- 230000002087 whitening effect Effects 0.000 claims abstract description 5
- 235000008216 herbs Nutrition 0.000 claims description 37
- 238000012549 training Methods 0.000 claims description 30
- 238000001514 detection method Methods 0.000 claims description 18
- 241000411851 herbal medicine Species 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 230000009193 crawling Effects 0.000 claims description 7
- 230000009471 action Effects 0.000 claims description 6
- 229940079593 drug Drugs 0.000 claims description 5
- 230000001939 inductive effect Effects 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 206010011224 Cough Diseases 0.000 claims description 3
- 206010062717 Increased upper airway secretion Diseases 0.000 claims description 3
- 208000004880 Polyuria Diseases 0.000 claims description 3
- XTXRWKRVRITETP-UHFFFAOYSA-N Vinyl acetate Chemical compound CC(=O)OC=C XTXRWKRVRITETP-UHFFFAOYSA-N 0.000 claims description 3
- 239000000428 dust Substances 0.000 claims description 3
- 239000000839 emulsion Substances 0.000 claims description 3
- 239000011521 glass Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 208000026435 phlegm Diseases 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 208000011580 syndromic disease Diseases 0.000 claims description 3
- 230000002936 tranquilizing effect Effects 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000010792 warming Methods 0.000 claims description 3
- 230000000507 anthelmentic effect Effects 0.000 claims description 2
- 239000003212 astringent agent Substances 0.000 claims description 2
- 230000002439 hemostatic effect Effects 0.000 claims description 2
- 239000008141 laxative Substances 0.000 claims description 2
- 230000001543 purgative effect Effects 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims description 2
- 229940124339 anthelmintic agent Drugs 0.000 claims 1
- 239000000921 anthelmintic agent Substances 0.000 claims 1
- 230000001680 brushing effect Effects 0.000 claims 1
- 239000002895 emetic Substances 0.000 claims 1
- NQBKFULMFQMZBE-UHFFFAOYSA-N n-bz-3-benzanthronylpyrazolanthron Chemical compound C12=CC=CC(C(=O)C=3C4=CC=CC=3)=C2C4=NN1C1=CC=C2C3=C1C1=CC=CC=C1C(=O)C3=CC=C2 NQBKFULMFQMZBE-UHFFFAOYSA-N 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 6
- 238000013499 data model Methods 0.000 abstract description 2
- 238000011160 research Methods 0.000 abstract description 2
- 238000011161 development Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000001737 promoting effect Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 230000036541 health Effects 0.000 description 3
- 229940126680 traditional chinese medicines Drugs 0.000 description 3
- 206010047700 Vomiting Diseases 0.000 description 2
- 208000006673 asthma Diseases 0.000 description 2
- 230000035619 diuresis Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000002203 pretreatment Methods 0.000 description 2
- 241000894007 species Species 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 210000005013 brain tissue Anatomy 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001914 calming effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000003930 cognitive ability Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000004925 denaturation Methods 0.000 description 1
- 230000036425 denaturation Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000029087 digestion Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000001678 irradiating effect Effects 0.000 description 1
- 210000004185 liver Anatomy 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- -1 slices Substances 0.000 description 1
- 238000010408 sweeping Methods 0.000 description 1
- 238000004809 thin layer chromatography Methods 0.000 description 1
- 230000001256 tonic effect Effects 0.000 description 1
- 230000008673 vomiting Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/55—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a Chinese medicinal material class identification method based on deep learning, which comprises two parts of establishing a learning model and identifying Chinese medicinal material decoction pieces. According to the method, after a data set is subjected to data enhancement by using a web crawler and an offline acquisition mode, an identification data model is established by a convolutional neural network, a traditional Chinese medicine decoction piece is subjected to sample processing and continuous image acquisition before the process of identifying the traditional Chinese medicine decoction piece, the acquired image is subjected to image fusion based on the convolutional neural network by using homomorphic filtering, after the image is processed by ZCA whitening, a Transformer is used for extracting characteristics, and the image characteristics are introduced into the convolutional neural network model for characteristic comparison, so that the categories of the sample traditional Chinese medicine are rapidly acquired, the number of factors influencing the identification accuracy in the processing process is greatly discharged, the identification rate is improved, theoretical support can be provided for the research of a traditional Chinese medicine rapid identification method, and the method has a very profound significance for emphasizing the modernization of the traditional Chinese medicine.
Description
Technical Field
The invention relates to the technical field of traditional Chinese medicine identification, in particular to a traditional Chinese medicine class identification method based on deep learning.
Background
At present, the identification of herbal pieces is basically determined by experts based on knowledge and experience, or by comparison of pictures. People in the post epidemic period have a relatively high enthusiasm for preventing traditional Chinese medicine health diseases, but most people are not professionals in the aspect, the resolution capability of traditional Chinese medicine decoction pieces is limited, the traditional Chinese medicines are various in variety and not completely standardized in the market, and a lot of related personnel in charge of purchasing cannot completely and accurately identify the traditional Chinese medicines.
Today with the development of computer technology, the identification of the types of Chinese medicinal decoction pieces can be easily realized by combining deep learning with a large amount of data and deploying the data in small programs, APP, websites and the like. The traditional Chinese medicine decoction pieces can improve the cognitive ability of people on the traditional Chinese medicine, and broaden the knowledge in the aspect of health preservation, so that the health condition can be known more, and in short, the life quality of people can be improved. Meanwhile, the development of traditional Chinese medicine and modern computer technology is promoted, the method is a inheritance and innovation for traditional Chinese medicine, and has very important significance for promoting the modernization of traditional Chinese medicine.
In the prior art, as Chinese patent numbers are: CN 105891172A, "a method for detecting and identifying different species and confusing traditional Chinese medicinal materials or traditional Chinese medicine decoction pieces", which comprises the following steps: the ultraviolet rays are applied to detection of the traditional Chinese medicinal materials, the ultraviolet rays are used for irradiating sections, powder, slices, solution extracting solutions or thin-layer chromatography development separators of the traditional Chinese medicinal materials or traditional Chinese medicinal decoction pieces, and the purposes of identifying different species and easily confusing the truth and the falseness of the traditional Chinese medicinal materials or the traditional Chinese medicinal decoction pieces are achieved through color change and color characteristics of fluorescence generated by irradiation.
However, in the prior art, the traditional Chinese medicinal material identification is realized by a physical/chemical characteristic extraction method, the pretreatment method is complicated, the identification rate is limited due to excessive factors influencing the identification accuracy in the treatment process, and the traditional Chinese medicinal material identification based on computer vision is less in development due to various factors such as shortage of resources, irregularity and complex real scenes of the traditional Chinese medicinal material data set, has uncertainty, increases the false detection rate and is harmful to the life safety of people.
Therefore, a method for identifying Chinese medicinal material classes based on deep learning is provided so as to solve the problems.
Disclosure of Invention
The invention aims to provide a traditional Chinese medicine material class identification method based on deep learning, and aims to solve the problems that the traditional Chinese medicine material identification is realized by a physical/chemical characteristic extraction method in the background technology, the pretreatment method is complicated, the identification rate is limited due to excessive factors influencing the identification accuracy in the treatment process, and the traditional Chinese medicine material identification based on computer vision is less in development due to various factors of shortage of traditional Chinese medicine material data set resources, irregularity and complex real scenes, has uncertainty, increases the false detection rate and is harmful to the life safety of people.
In order to achieve the purpose, the invention provides the following technical scheme: a traditional Chinese medicine class identification method based on deep learning comprises two parts of establishing a learning model and identifying traditional Chinese medicine decoction pieces, wherein the step of establishing the learning model comprises the following steps:
s10, classifying the Chinese herbal medicine decoction pieces: marking the types of the traditional Chinese medicinal materials by adopting a function classification method, and according to the commonness of the same type of medicines in the aspects of medicine property, compatibility and contraindication, utilizing the similarities and the differences of the similar types of medicines according to the strength of action and different action parts;
s11, data set acquisition: multithread crawling is carried out on the Baidu picture on a Scapy frame by using a Python crawler, high-definition cameras are used in an offline trading market to shoot and collect traditional Chinese medicine data, and label annotation is carried out on the collected traditional Chinese medicine data according to the traditional Chinese medicine types in the step S10;
s12, preprocessing of the data set: removing repeated data in RGB image data of the traditional Chinese medicinal materials according to the types of the traditional Chinese medicinal materials, dividing the data into a training set, a verification set and a test set according to the ratio of 6:2:2, then performing data enhancement, and establishing a sample label array by adopting a 2D one-hot coding label;
s13, model training: selecting Alex Net, Goog LeNet and Squeeze Net as basic structures, initializing model parameters by using parameters obtained by training Alex Net, Squeeze Net and Goog LeNet on an Image Net data set, and then carrying out fine tuning training;
the identification of the traditional Chinese medicine decoction pieces comprises the following steps:
s20, preparing a detection sample;
s21, acquiring and arranging the detection sample image;
and S22, importing the detection sample image obtained in the step S22 into the convolution neural network model for convolution processing.
Preferably, in step S10, the Chinese medicinal herbs include 19 kinds of herbs selected from the group consisting of exterior syndrome-relieving herbs, heat-clearing herbs, purgative herbs, wind-damp-dispelling herbs, dampness-resolving herbs, diuresis-inducing and dampness-excreting herbs, interior-warming herbs, qi-regulating herbs, digestion-promoting herbs, insect-expelling herbs, bleeding-stopping herbs, blood-activating herbs, phlegm-resolving, cough-relieving and asthma-relieving herbs, tranquilizing herbs, liver-calming and wind-extinguishing herbs, resuscitation-inducing herbs, tonifying herbs, astringents, and vomiting-promoting herbs.
Preferably, in step S12, the data enhancement is to randomly rotate the images of the traditional Chinese medicinal materials by 30 °, randomly shift the images by 20% in the horizontal direction and the vertical direction, randomly shift the intensity of the miscut transform by 0.2, set the amplitude of the random scaling of the images to 0.2, and adjust all the images of the traditional Chinese medicinal materials to 150 × 150 pixels after the images are randomly horizontally flipped.
Preferably, in step S13, the model training includes the following steps:
s130, dividing a sub-training set based on a Bagging method;
s131, training by utilizing each sub-training set according to a feature fusion network training mode to obtain a plurality of weak classifiers;
and S132, integrating the weak classifiers into a strong classifier.
Preferably, in step S20, the step of preparing the test sample is to sweep away fine dust on the surface of the herbal pieces with a brush, and the herbal pieces are fixed on a glass slide with a vinyl acetate emulsion.
Preferably, in step S21, the acquiring and processing of the detection sample image includes:
s220, adjusting the distance between the traditional Chinese medicine decoction pieces in the step S20 to the edge by using an electronic eyepiece;
s221, gradually adjusting the focal length to the center and continuously shooting and collecting until the whole image is shot;
s222, after the resolutions of the images acquired in step S221 are unified to 28 × 28 pixels, the redundant images are deleted.
Preferably, in step S22, the method includes the following steps:
s220, carrying out image fusion on the acquired image and the convolution-based neural network by using homomorphic filtering;
s221, processing the image processed in the step S220 through ZCA whitening;
s222, extracting features by using a Transformer;
and S223, performing identification through a Softmax classifier.
Preferably, in step S220, the image focusing degree is determined by performing cushion detection based on the convolutional neural network model, the obtained focus is converted into a binary map by determining a threshold, and the image is fused by a weighted average method after the binary map is extracted by taking out a small region and guiding a filtering optimization mean filter.
Compared with the prior art, the invention has the beneficial effects that:
according to the method, after a data set is subjected to data enhancement by using a web crawler and an offline acquisition mode, an identification data model is established by a convolutional neural network, a traditional Chinese medicine decoction piece is subjected to sample processing and continuous image acquisition before the process of identifying the traditional Chinese medicine decoction piece, the acquired image is subjected to image fusion based on the convolutional neural network by using homomorphic filtering, after the image is processed by ZCA whitening, a Transformer is used for extracting characteristics, and the image characteristics are introduced into the convolutional neural network model for characteristic comparison, so that the categories of the sample traditional Chinese medicine are rapidly acquired, the number of factors influencing the identification accuracy in the processing process is greatly discharged, the identification rate is improved, theoretical support can be provided for the research of a traditional Chinese medicine rapid identification method, and the method has a very profound significance for emphasizing the modernization of the traditional Chinese medicine.
Drawings
FIG. 1 is a flow chart of a method for identifying Chinese medicinal material categories based on deep learning according to the present invention;
FIG. 2 is a Scapy crawling flowchart of a Chinese medicinal material class identification method based on deep learning according to the present invention;
fig. 3 is a schematic diagram of a standard neural network of the method for identifying Chinese medicinal material classes based on deep learning.
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.
Referring to fig. 1-3, the present invention provides a technical solution: a Chinese medicinal material class identification method based on deep learning. Among them, the convolutional neural network is a feedforward neural network, which includes convolutional calculation and has a deep structure, and thus is one of representative algorithms of deep learning. With the continuous progress of science and technology, people are inspired to create a neural network when researching human brain tissues. The neural network consists of a plurality of mutually-connected neurons, and can enhance or inhibit signals among different neurons by adjusting and transmitting a weight coefficient x which is mutually connected.
Wherein, the learning model establishment comprises the following steps:
firstly, classifying Chinese medicinal decoction pieces: the categories of the traditional Chinese medicinal materials are various and complex, the total number of the traditional Chinese medicinal materials is about 8000 according to the statistical results of related workers in recent years, rare medicines are removed, and about 700 traditional Chinese medicinal materials are commonly used by people in daily life. In order to better study, study and apply the Chinese medicinal materials.
Marking the types of the Chinese medicinal materials by a function classification method, and performing group analogy according to the similarity of the same type of medicaments in aspects of property, compatibility and contraindication and according to the strength of action and the difference and the similarity of action parts among the same type of medicaments; the Chinese medicinal materials include 19 kinds of exterior syndrome relieving medicine, heat clearing medicine, cathartic medicine, rheumatism expelling medicine, dampness eliminating medicine, diuresis inducing medicine, interior warming medicine, qi regulating medicine, digestion promoting medicine, anthelmintic medicine, hemostatic medicine, blood circulation promoting medicine, phlegm eliminating, cough relieving, asthma relieving medicine, tranquilizing medicine, liver calming, wind extinguishing medicine, resuscitation inducing medicine, tonic medicine, astringing medicine and vomiting promoting medicine.
Secondly, data set acquisition: as the traditional Chinese medicinal materials are special Chinese medicinal materials, and no open-source traditional Chinese medicinal material data set exists on the network at present, the project uses a Python crawler to perform multithread crawling on hundred-degree pictures on a Scapy frame. The script has a crawler framework with powerful functions, high crawling efficiency, multiple related extension components, and high configurable and extensible degrees.
The crawler-based asynchronous processing framework based on the Twisted is a crawler framework realized by pure Python, has clear framework and low module coupling degree, can flexibly meet various requirements, and simply speaking, Scapy is an application framework written for crawling website data and extracting data, and the crawling flow is shown in FIG. 2. Meanwhile, the offline trading market uses a high-definition camera to shoot and collect the traditional Chinese medicine data, and labels are added to the collected traditional Chinese medicine data according to the types of the traditional Chinese medicines.
Thirdly, preprocessing a data set: removing repeated data in RGB image data of the traditional Chinese medicinal materials according to the types of the traditional Chinese medicinal materials, dividing the data into a training set, a verification set and a test set according to the ratio of 6:2:2, then performing data enhancement, and establishing a sample label array by adopting a 2D one-hot coding label; the data enhancement is that the traditional Chinese medicine images are randomly rotated by 30 degrees, randomly translated by 20 percent in the horizontal direction and the vertical direction, the random shearing transformation intensity is 0.2, the random scaling amplitude of the images is set to be 0.2, and after the images are randomly and horizontally turned, all the traditional Chinese medicine images are adjusted to be 150 multiplied by 150 pixels.
Fourthly, model training: a standard convolutional neural network is typically composed of an input layer, convolutional layer, pooling layer, fully-connected layer, and output layer, as shown in fig. 3. The first layer is an input layer with the size of 28x28, the first layer passes through a convolution layer with the size of 20x2424, the obtained result is input into a pooling layer, and the fourth layer, namely a full connection layer in the figure, is finally output.
Selecting Alex Net, Goog LeNet and Squeeze Net as a basic structure, initializing model parameters by using parameters obtained by training Alex Net, Squeeze Net and Goog LeNet on an Image Net data set, and then performing fine tuning training, wherein the fine tuning training content is as follows: firstly, the output node of a Softmax classifier in the Alex Net, Squeeze Net and Goog LeNet networks is modified to be 98, parameters obtained by Image Net training are used for initializing the parameters of the three networks, then the Alex Net, Squeeze Net and Goog LeNet networks are finely adjusted by using the medicinal material data, and a model with a better effect is trained. The network structure is then adjusted to fuse the underlying features with the higher-level features by adding a BN layer and a feature connection layer (concat layer), and the modified networks are called Alex Net-fusion, Squeeze Net-fusion, Goog LeNet-fusion. Finally, the network parameters obtained by training in the step 1) are assigned to the parameters of Alex Net-fusion, Squeeze Net-fusion and Goog Le Net-fusion, and then fine adjustment is continued on the Chinese medicinal material data set at a smaller learning rate.
The model training comprises the following steps:
1) dividing a sub-training set based on a Bagging method;
2) training by utilizing each sub-training set according to a feature fusion network training mode to obtain a plurality of weak classifiers;
3) and integrating each weak classifier into a strong classifier.
The identification of the traditional Chinese medicine decoction pieces comprises the following steps:
firstly, detecting sample preparation: the detection sample preparation comprises the steps of sweeping off fine dust on the surfaces of the traditional Chinese medicine decoction pieces by using a brush, and fixing the traditional Chinese medicine decoction pieces on a glass slide by using vinyl acetate emulsion.
Secondly, obtaining and arranging a detection sample image: firstly, the distance of the Chinese medicinal decoction pieces is adjusted to the edge by using an electronic eyepiece, then the focus is gradually adjusted to the center and shooting and collection are continuously carried out until the whole image is shot, and finally, after the resolution of the collected image is unified to 28 multiplied by 28 pixels, redundant images are deleted.
And thirdly, introducing the obtained detection sample image into a convolution neural network model for convolution processing. The method comprises the following steps:
1) carrying out image fusion on the acquired image and the convolution-based neural network by using homomorphic filtering; the method comprises the steps of judging the focusing degree of an image through cushion detection based on a convolutional neural network model, converting an obtained focus into a binary map through determining a threshold, taking out a small region and performing binary map through a guide filtering optimization mean filter, and fusing the image through a weighted average method.
2) The image is processed by ZCA whitening.
3) Features, CNN and Transformer, were extracted using Transformer. CNN is a layered data representation mode, and the feature representation of the high level depends on the feature representation of the bottom level, and the features with higher semantic information are abstractly extracted from shallow to deep step by step. In addition, CNN also has some translational invariance and translational denaturation. The receptive field of the CNN is small, and theoretically, the receptive field can be very large by stacking CNN networks, but the actual receptive field is much smaller than the theoretical one. And the Transformer can establish a global dependency. Feature extraction by the Transformer is better able to exploit the large amount of data, which is demonstrated on the Natural Language (NLP) task, over a large set of data, the Transformer exceeds the best classification network of CNN structures.
4) The discrimination was performed by Softmax classifier.
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes in the embodiments and/or modifications of the invention can be made, and equivalents and modifications of some features of the invention can be made without departing from the spirit and scope of the invention.
Claims (8)
1. A traditional Chinese medicine class identification method based on deep learning is characterized by comprising two parts of establishing a learning model and identifying traditional Chinese medicine decoction pieces, wherein the establishing of the learning model comprises the following steps:
s10, classifying the Chinese herbal medicine decoction pieces: marking the types of the traditional Chinese medicinal materials by adopting a function classification method, and according to the commonness of the same type of medicines in the aspects of medicine property, compatibility and contraindication, utilizing the similarities and the differences of the similar types of medicines according to the strength of action and different action parts;
s11, data set acquisition: multithread crawling is carried out on Baidu pictures on a script frame by using a Pyron crawler, high-definition cameras are used in an offline trading market for shooting and collecting traditional Chinese medicine data, and label annotation is carried out on the collected traditional Chinese medicine data according to the traditional Chinese medicine types in the step S10;
s12, preprocessing of the data set: removing repeated data in RGB image data of the traditional Chinese medicinal materials according to the types of the traditional Chinese medicinal materials, dividing the data into a training set, a verification set and a test set according to the ratio of 6:2:2, then performing data enhancement, and establishing a sample label array by adopting a 2D one-hot coding label;
s13, model training: selecting Alex Net, Goog LeNet and Squeeze Net as basic structures, initializing model parameters by using parameters obtained by training Alex Net, Squeeze Net and Goog LeNet on an Image Net data set, and then carrying out fine tuning training;
the identification of the traditional Chinese medicine decoction pieces comprises the following steps:
s20, preparing a detection sample;
s21, acquiring and arranging the detection sample image;
and S22, importing the detection sample image obtained in the step S22 into the convolutional neural network model for convolution processing.
2. The method for identifying Chinese herbal medicines based on deep learning of claim 1, wherein in step S10, the Chinese herbal medicines include 19 kinds of herbs selected from the group consisting of exterior-syndrome-relieving herbs, heat-clearing herbs, purgative herbs, rheumatism-expelling herbs, dampness-resolving herbs, diuresis-promoting and dampness-excreting herbs, interior-warming herbs, qi-regulating herbs, digestion-promoting herbs, anthelmintics, hemostatic herbs, blood-activating herbs, phlegm-resolving and cough-relieving herbs, tranquilizing herbs, liver-calming and wind-extinguishing herbs, resuscitation-inducing herbs, tonifying herbs, astringents, and emetics.
3. The method for identifying Chinese herbal medicine products based on deep learning as claimed in claim 1, wherein in step S12, the data enhancement is to randomly rotate the Chinese herbal medicine images by 30 °, randomly shift the images by 20% in horizontal and vertical directions, randomly shift the intensity of the cross-cut transformation by 0.2, set the amplitude of the random scaling of the images to 0.2, and adjust all the Chinese herbal medicine images to 150 × 150 pixels after the images are randomly horizontally flipped.
4. The method for identifying Chinese medicinal material classes based on deep learning of claim 1, wherein in step S13, the model training comprises the following steps:
s130, dividing a sub-training set based on a Bagging method;
s131, training by utilizing each sub-training set according to a feature fusion network training mode to obtain a plurality of weak classifiers;
and S132, integrating the weak classifiers into a strong classifier.
5. The method for identifying Chinese medicinal material categories based on deep learning as claimed in claim 1, wherein in step S20, the detection sample is prepared by brushing off fine dust on the surface of Chinese medicinal material decoction pieces, and fixing the Chinese medicinal material decoction pieces on a glass slide by using vinyl acetate emulsion.
6. The method for identifying traditional Chinese medicine products based on deep learning of claim 5, wherein in step S21, the acquiring and sorting of the detection sample images comprises the following steps:
s220, adjusting the distance between the Chinese herbal medicine decoction pieces in the step S20 to the edge by using an electronic eyepiece;
s221, gradually adjusting the focal length to the center and continuously shooting and collecting until the whole image is shot;
s222, after the resolutions of the images acquired in step S221 are unified to 28 × 28 pixels, the redundant images are deleted.
7. The method for identifying traditional Chinese medicinal materials based on deep learning of claim 1, wherein in step S22, the method comprises the following steps:
s220, carrying out image fusion on the acquired image and the convolution-based neural network by using homomorphic filtering;
s221, processing the image processed in the step S220 through ZCA whitening;
s222, extracting features by using a Transformer;
and S223, performing identification through a Softmax classifier.
8. The method for identifying traditional Chinese medicine products based on deep learning of claim 7, wherein in step S220, the gel pad detection is performed through the convolutional neural network model to determine the focusing degree of the images, the obtained focus is converted into a binary map through determining a threshold, and the images are fused by a weighted average method after the binary map is extracted and subjected to guided filtering optimization mean filter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210412607.5A CN114998639B (en) | 2022-04-19 | 2022-04-19 | Deep learning-based traditional Chinese medicine category identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210412607.5A CN114998639B (en) | 2022-04-19 | 2022-04-19 | Deep learning-based traditional Chinese medicine category identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114998639A true CN114998639A (en) | 2022-09-02 |
| CN114998639B CN114998639B (en) | 2024-04-26 |
Family
ID=83025847
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210412607.5A Active CN114998639B (en) | 2022-04-19 | 2022-04-19 | Deep learning-based traditional Chinese medicine category identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114998639B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117011612A (en) * | 2023-08-16 | 2023-11-07 | 海南省新超豪信息技术有限公司 | AI identification method for traditional Chinese medicinal materials |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107958257A (en) * | 2017-10-11 | 2018-04-24 | 华南理工大学 | A kind of Chinese traditional medicinal materials recognition method based on deep neural network |
| CN109190643A (en) * | 2018-09-14 | 2019-01-11 | 华东交通大学 | Based on the recognition methods of convolutional neural networks Chinese medicine and electronic equipment |
| CN109615010A (en) * | 2018-12-13 | 2019-04-12 | 济南大学 | Chinese traditional medicinal materials recognition method and system based on double scale convolutional neural networks |
| US20190114511A1 (en) * | 2017-10-16 | 2019-04-18 | Illumina, Inc. | Deep Learning-Based Techniques for Training Deep Convolutional Neural Networks |
| CN111598126A (en) * | 2020-04-08 | 2020-08-28 | 天津大学 | Lightweight traditional Chinese medicinal material identification method |
| CN113159044A (en) * | 2021-04-01 | 2021-07-23 | 中国公路工程咨询集团有限公司 | Deep learning-based road material identification method for convolutional neural network |
| US20220068501A1 (en) * | 2020-09-01 | 2022-03-03 | Autonomous Healthcare, Inc. | System and method for tracking sensitive medications |
-
2022
- 2022-04-19 CN CN202210412607.5A patent/CN114998639B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107958257A (en) * | 2017-10-11 | 2018-04-24 | 华南理工大学 | A kind of Chinese traditional medicinal materials recognition method based on deep neural network |
| US20190114511A1 (en) * | 2017-10-16 | 2019-04-18 | Illumina, Inc. | Deep Learning-Based Techniques for Training Deep Convolutional Neural Networks |
| CN109190643A (en) * | 2018-09-14 | 2019-01-11 | 华东交通大学 | Based on the recognition methods of convolutional neural networks Chinese medicine and electronic equipment |
| CN109615010A (en) * | 2018-12-13 | 2019-04-12 | 济南大学 | Chinese traditional medicinal materials recognition method and system based on double scale convolutional neural networks |
| CN111598126A (en) * | 2020-04-08 | 2020-08-28 | 天津大学 | Lightweight traditional Chinese medicinal material identification method |
| US20220068501A1 (en) * | 2020-09-01 | 2022-03-03 | Autonomous Healthcare, Inc. | System and method for tracking sensitive medications |
| CN113159044A (en) * | 2021-04-01 | 2021-07-23 | 中国公路工程咨询集团有限公司 | Deep learning-based road material identification method for convolutional neural network |
Non-Patent Citations (2)
| Title |
|---|
| 吴雪;宋晓茹;高嵩;陈超波;: "基于数据增强的卷积神经网络火灾识别", 科学技术与工程, no. 03, 28 January 2020 (2020-01-28) * |
| 程曦;吴云志;张友华;乐毅;: "基于深度卷积神经网络的储粮害虫图像识别", 中国农学通报, no. 01, 3 January 2018 (2018-01-03) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117011612A (en) * | 2023-08-16 | 2023-11-07 | 海南省新超豪信息技术有限公司 | AI identification method for traditional Chinese medicinal materials |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114998639B (en) | 2024-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107958244A (en) | A kind of face identification method and device based on the fusion of video multiframe face characteristic | |
| CN110598535B (en) | Face recognition analysis method used in monitoring video data | |
| CN107133575A (en) | A kind of monitor video pedestrian recognition methods again based on space-time characteristic | |
| Eberhardt et al. | How deep is the feature analysis underlying rapid visual categorization? | |
| CN108830818A (en) | A kind of quick multi-focus image fusing method | |
| CN106897673A (en) | A kind of recognition methods again of the pedestrian based on retinex algorithms and convolutional neural networks | |
| CN108776774A (en) | A kind of human facial expression recognition method based on complexity categorization of perception algorithm | |
| CN101999900A (en) | Living body detecting method and system applied to human face recognition | |
| CN112784763A (en) | Expression recognition method and system based on local and overall feature adaptive fusion | |
| CN105574509A (en) | Face identification system playback attack detection method and application based on illumination | |
| CN113069080A (en) | Difficult airway assessment method and device based on artificial intelligence | |
| CN109815945A (en) | A kind of respiratory tract inspection result interpreting system and method based on image recognition | |
| CN110163567A (en) | Classroom roll calling system based on multitask concatenated convolutional neural network | |
| CN108596174A (en) | A kind of lesion localization method of skin disease image | |
| Asare et al. | Iron deficiency anemia detection using machine learning models: A comparative study of fingernails, palm and conjunctiva of the eye images | |
| CN112151167A (en) | Intelligent screening method for six-age dental caries of children based on deep learning | |
| CN114998639A (en) | Chinese medicinal material class identification method based on deep learning | |
| CN114998303A (en) | Small intestine interstitial tumor detection method with strong feature extraction capability | |
| CN112036520A (en) | Panda age identification method and device based on deep learning and storage medium | |
| CN113221655A (en) | Face spoofing detection method based on feature space constraint | |
| Dong et al. | Semi-supervised domain alignment learning for single image dehazing | |
| CN109902730A (en) | Broken strand of power transmission line detection method based on deep learning | |
| Duffner et al. | A neural scheme for robust detection of transparent logos in TV programs | |
| CN112906510A (en) | Fishery resource statistical method and system | |
| CN115760810B (en) | Medical image segmentation apparatus, method and computer-readable storage medium |
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 |