CN108229979A - Based on wechat small routine and the American Ginseng of depth learning technology tasting assistant - Google Patents
Based on wechat small routine and the American Ginseng of depth learning technology tasting assistant Download PDFInfo
- Publication number
- CN108229979A CN108229979A CN201810066788.4A CN201810066788A CN108229979A CN 108229979 A CN108229979 A CN 108229979A CN 201810066788 A CN201810066788 A CN 201810066788A CN 108229979 A CN108229979 A CN 108229979A
- Authority
- CN
- China
- Prior art keywords
- american ginseng
- tasting
- layer
- assistant
- small routine
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of based on wechat small routine and the American Ginseng of depth learning technology tasting assistant, include the mobile equipment for being equipped with wechat client, American Ginseng tasting assistant's small routine terminal and cloud server analysis platform；Mobile equipment uses wechat client by network connection, scan the two-dimensional code load operating American Ginseng tasting assistant's small routine terminal, American Ginseng tasting assistant's small routine terminal calls the camera of mobile equipment to shoot American Ginseng sample image to be identified, and it is uploaded to cloud server analysis platform, cloud server analysis platform carries out level identification using American Ginseng tasting deep learning neural network model to American Ginseng sample image to be identified, and by qualification result back to mobile terminal, and analysis result is shown in American Ginseng tasting assistant's small routine terminal.The present invention realize American Ginseng quality identification portability, it is rapid, objectify, solve the problems, such as amateur purchasing group when purchasing medicine without tool it is available, without according to can be according to.
Description
Technical field
It is specifically a kind of to be based on wechat small routine and deep learning the present invention relates to rare Chinese medicine quality identification technical field
The American Ginseng tasting assistant of technology.
Background technology
American Ginseng is quite active, frequent in health products trade marketing.Some businessmans are in order to obtain the illegal profit for great number
The phenomenon that moistening, adulterating, mix the spurious with the genuine is extremely serious, has upset market order.The method for being currently used in TCD identificafion is main
Have experience discriminating, microscopical characters, physico-chemical analysis differentiate, spectral imaging analysis differentiate etc., these methods all to appraiser from
Industry experience, professional habit have stronger dependence, it is desirable that and appraiser has quite deep theoretical foundation and abundant smell of powder,
And microscopical characters, physico-chemical analysis differentiate, spectral imaging analysis authentication equipment not Portable belt, cause amateur Gou Yao groups can without tool
With, without according to can be according to.
Invention content
It is an object of the invention to provide one kind based on wechat small routine and depth for the amateur purchasing group of American Ginseng
The American Ginseng tasting assistant of habit technology, realize American Ginseng quality identification portability, it is rapid, objectify, solve amateur purchase
Group when purchasing medicine without tool it is available, without according to can according to the problem of.
Technical scheme is as follows：
It is a kind of that assistant is judged based on wechat small routine and the American Ginseng of deep learning technology, it is characterised in that：Include and be equipped with
Mobile equipment, American Ginseng tasting assistant's small routine terminal and the cloud server analysis platform of wechat client；The movement
Equipment, using wechat client, it is whole to scan the two-dimensional code American Ginseng tasting assistant small routine described in load operating by network connection
End, American Ginseng tasting assistant's small routine terminal call the camera of the mobile equipment to shoot American Ginseng sample to be identified
Image, and the cloud server analysis platform is uploaded to, the cloud server analysis platform judges depth using American Ginseng
Learning neural network model carries out American Ginseng sample image to be identified level identification, and qualification result is returned to the shifting
Dynamic terminal, and show analysis result in American Ginseng tasting assistant's small routine terminal.
Described judges assistant based on wechat small routine and the American Ginseng of deep learning technology, it is characterised in that：Described
The interface of American Ginseng tasting assistant's small routine terminal is made of shoot button, preview picture area and results display area；Wherein, as a result
Viewing area can list 4 graded categories and two parameters of confidence level, and be arranged from high to low by confidence level.
Described judges assistant based on wechat small routine and the American Ginseng of deep learning technology, it is characterised in that：Described
Cloud server analysis platform includes image and receives and preprocessing module, deep neural network analysis module and result loopback mould
Block；The input number that described image is received with the output data of preprocessing module is the deep learning neural network analysis module
According to, wherein, described image receives the American Ginseng sample to be identified for being used to receive the temporary mobile equipment upload with preprocessing module
Simultaneously preceding pretreatment is identified in this image, wherein the content pre-processed includes：It is sized, data channel exchanges, Color Channel
Exchange；The result loopback module is used to extract the analysis result data of the deep learning neural network analysis module,
And it arranges and transfers to American Ginseng tasting assistant's small routine terminal through wechat public platform server for JSON forms.
Described judges assistant based on wechat small routine and the American Ginseng of deep learning technology, it is characterised in that：Described
American Ginseng judges the preparation method such as following steps of deep learning neural network model：
American Ginseng is divided into first, second and third graded category, and by above three by step 1 according to American Ginseng commercial specification index
Graded category is collected and mark American Ginseng sample image, and collects American Ginseng Its Common Confused Varieties image, and Its Common Confused Varieties are marked
For fourth estate classification, image data base is established, wherein, the image data of each graded category is not less than 1000 width；
The pretreatment of step 2, image data sample, including：By image using mirror image rotation, scaling, lighting simulation to picture number
Sample size amplification is realized according to library；256*256 sizes are uniformly scaled, and are converted into LMDB formatted datas；
Step 3, establishment are upset by the label text file of " image path and class label " form and are ranked sequentially label text text
" image path and class label " record strip in part, to the record strip of each graded category, by 5：1 ratio extracts training set
With test set；
Step 4, using Caffe deep learning frames, according to the parameter initialization pre-training neural network of AlexNet models, and
The output layer neuron number of classification is changed to 4；Tuning is carried out, and protect to network parameter using the training set that step 3 obtains
It deposits, export American Ginseng tasting deep learning neural network model；
Step 5, beyond the clouds arrangement step 4 obtain American Ginseng tasting deep learning neural network model.
Described judges assistant based on wechat small routine and the American Ginseng of deep learning technology, it is characterised in that：Described
American Ginseng tasting deep learning neural network model includes：Input layer, convolutional layer 1,2,3,4,5, pond layer 1,2,3, connect layer entirely
1st, 2 and Softmax output layers；The input layer, convolutional layer 1, pond layer 1, convolutional layer 2, pond layer 2, convolutional layer 3, convolution
Layer 4, convolutional layer 5, pond layer 3, it is complete connect layer 1, the full layer 2 that connects sequentially is connected with Softmax output layers, the output data of upper level is
The input data of next stage；Each input picture of the input layer divides variability size to divide tri- face of RGB for 227*227
Color dimension inputs；The convolutional layer 1, convolution nuclear volume are 96, and the size of each convolution kernel is 11*11, and convolution stride is
4；The pond layer 1, using maximization pond mode, Qi Chihua ranging from 3*3, sampling step length 2；The convolutional layer 2,
Its convolution nuclear volume is 256, and the size of each convolution kernel is 5*5, and convolution group is 2, and it is 2 to expand edge；The pond layer 2,
Using maximization pond mode, Qi Chihua ranging from 3*3, sampling step length 2；The convolutional layer 3, convolution nuclear volume are
384, the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；The convolutional layer 4, convolution nuclear volume are
384, the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；The convolutional layer 5, convolution nuclear volume are
256, the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；The pond layer 3, using maximization pond
Mode, Qi Chihua ranging from 3*3, sampling step length 2；Described connects layer 1 entirely, and the wherein neuron number of this layer is 4096,
By the way of stochastic gradient descent；Described connects layer 2 entirely, and the wherein neuron number of this layer is 4096, using boarding steps
Spend the mode declined；The Softmax output layers, using Softmax return mode, neuron number be 4, i.e., with the West
4 graded categories of ginseng are corresponding, and the value of output belongs to 4 graded category American Ginseng figures for American Ginseng sample image to be identified
The probability value of picture.
Beneficial effects of the present invention：
The present invention provides a kind of American Ginseng based on wechat small routine and deep learning technology for the amateur purchasing group of American Ginseng
Judge assistant, realize American Ginseng quality identification portability, it is rapid, objectify, solve amateur purchasing group purchase medicine
When without tool it is available, without according to can according to the problem of.
Description of the drawings
Fig. 1 is present system structure diagram.
Fig. 2 is the surface chart that American Ginseng judges assistant's small routine terminal in the present invention.
Fig. 3 is deep learning neural metwork training, identification classification process figure in the present invention.
Fig. 4 judges deep learning neural network model composition schematic diagram for American Ginseng in the present invention.
Specific embodiment
Technical solution and implementation process in order to further illustrate the present invention carry out detailed with reference to Fig. 1, Fig. 2, Fig. 3, Fig. 4
Explanation.
It is of the invention a kind of based on wechat small routine and the American Ginseng of deep learning technology tasting assistant referring to Fig. 1, Fig. 2,
Include the Android mobile phone terminal 5 for being equipped with wechat client, American Ginseng tasting assistant's small routine terminal 1 and cloud server point
Analyse platform 7；Android mobile phone terminal 5, using wechat client, scans the two-dimensional code the tasting of load operating American Ginseng by network connection
Assistant's small routine terminal 1, American Ginseng tasting assistant's small routine terminal 1 call the camera shooting of Android mobile phone terminal 5 to be identified
American Ginseng sample image, and be uploaded to cloud server analysis platform 7, cloud server analysis platform 7 uses American Ginseng product
The deep learning neural network model that reflects carries out level identification, and qualification result is returned to American Ginseng sample image to be identified
Android mobile phone terminal 5, and show analysis result in American Ginseng tasting assistant's small routine terminal 1.
The interface of American Ginseng tasting assistant's small routine terminal 1 is by shoot button 2, preview picture area 3,4 groups of results display area
Into；Wherein results display area 4 can list 4 graded categories and two parameters of confidence level, and be arranged from high to low by confidence level.Cloud
End server analysis platform 7 includes image and receives and preprocessing module 8, deep neural network analysis module 9 and result loopback
Module 10；It is the input data of deep neural network analysis module 9 that image, which is received with the output data of preprocessing module 8,；Wherein,
Image is received goes forward side by side with preprocessing module 8 for receiving the American Ginseng sample image to be identified that temporary Android mobile phone terminal 5 uploads
It is pre-processed before row identification, wherein the content pre-processed includes：It is sized, data channel exchanges, the exchange of Color Channel；As a result
It returns to module 10 to be used to extract the analysis result data of deep neural network analysis module 10, and it is JSON forms through wechat to arrange
Public platform server 6 transfers to American Ginseng tasting assistant's small routine terminal 1.
Referring to Fig. 1, Fig. 2, specific embodiment：The client of certain the purchase American Ginseng Android mobile phone for being equipped with wechat client
Terminal 5, using wechat client, it is whole to scan the two-dimensional code load operating American Ginseng tasting assistant small routine by wireless network connection
Button 2 of taking pictures on 1 interface of American Ginseng tasting assistant's small routine terminal, the camera shooting for calling Android mobile phone terminal 5 included are clicked in end 1
Head shoots American Ginseng sample image to be identified, and servicing transfer by network Http through wechat public platform server 6 is uploaded to cloud
Server analysis platform 7 is held, cloud server analysis platform 7 receives temporary to be identified with the reception of preprocessing module 8 by image
American Ginseng sample image, and size adjustment is carried out to American Ginseng sample image（It is scaled：256*256）, data channel exchange, face
The exchange pretreatment of chrominance channel, pretreated image data are input to the deep neural network of cloud server analysis platform 7
Analysis is identified in analysis module 9, and analysis result is transmitted to cloud service by deep neural network analysis module 9 after the completion of analysis
The result loopback module 10 of device analysis platform 7；As a result after analysis result is encoded to JSON forms by loopback module 10, pass through network
American Ginseng tasting assistant's small routine terminal 1, American Ginseng tasting assistant's small routine terminal are relayed to through wechat public platform server 6
1 is presented in result by the form of " graded category+confidence level " results display area 4 of American Ginseng tasting assistant's small routine terminal 1.
Referring to Fig. 3, of the invention is a kind of based on wechat small routine and the American Ginseng of deep learning technology tasting assistant, the West
The preparation method such as following steps of ginseng tasting deep learning neural network model：
American Ginseng is divided into first, second and third graded category, and by above three by step 1 according to American Ginseng commercial specification index
Graded category is collected and mark American Ginseng sample image, and collects American Ginseng Its Common Confused Varieties image, and Its Common Confused Varieties are marked
For fourth estate classification, image data base is established, wherein, the image data of each graded category is not less than 1000 width；
The pretreatment of step 2, image data sample, including：By image using mirror image rotation, scaling, lighting simulation to picture number
Sample size amplification is realized according to library；256*256 sizes are uniformly scaled, and are converted into LMDB formatted datas；
Step 3, establishment are upset by the label text file of " image path and class label " form and are ranked sequentially label text text
" image path and class label " record strip in part, to the record strip of each graded category, by 5：1 ratio extracts training set
With test set；
Step 4, using Caffe deep learning frames, according to the parameter initialization pre-training neural network of AlexNet models, and
The output layer neuron number of classification is changed to 4；Tuning is carried out, and protect to network parameter using the training set that step 3 obtains
It deposits, export American Ginseng tasting deep learning neural network model；
Step 5, beyond the clouds arrangement step 4 obtain American Ginseng tasting deep learning neural network model.
Referring to Fig. 4, of the invention is a kind of based on wechat small routine and the American Ginseng of deep learning technology tasting assistant, the West
Ginseng tasting deep learning neural network model includes：Input layer I1,1 C1 of convolutional layer, 2 C2 of convolutional layer, 3 C3 of convolutional layer, convolution
Layer 4 C4,5 C5 of convolutional layer, 1 P1 of pond layer, 2 P2 of pond layer, 3 P3 of pond layer, entirely connect 1 F1 of layer, entirely connect layer 2F2 and
Softmax output layers O1；Input layer I1,1 C1 of convolutional layer, 1 P1 of pond layer, 2 C2 of convolutional layer, 2 P2 of pond layer, convolutional layer 3
C3,4 C4 of convolutional layer, 5 C5 of convolutional layer, 3 C3 of pond layer, complete connect that 1 F1 of layer, complete to connect 2 F2 and Softmax output layers O1 of layer suitable
Sequence is connected, and the output data of upper level is the input data of next stage；Each input picture of input layer I1 divides variability size
For 227*227, divide tri- color dimension inputs of RGB；1 C1 of convolutional layer, convolution nuclear volume are 96, the size of each convolution kernel
For 11*11, convolution stride is 4；1 P1 of pond layer, using pond mode is maximized, Qi Chihua ranging from 3*3, sampling step length is
2；Convolutional layer 2C2, convolution nuclear volume are 256, and the size of each convolution kernel is 5*5, and convolution group is 2, and it is 2 to expand edge；Pond
Change 2 P2 of layer, using maximization pond mode, Qi Chihua ranging from 3*3, sampling step length 2；Convolutional layer 3C3, convolution check figure
It is 384 to measure, and the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；Convolutional layer 4C4, convolution nuclear volume are
384, the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；Convolutional layer 5C5, convolution nuclear volume are 256,
The size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；Pond layer 3P3, using maximization pond mode, pond
Change ranging from 3*3, sampling step length 2；Complete to connect a layer 1F1, the wherein neuron number of this layer is 4096, using under stochastic gradient
The mode of drop；Complete to connect a layer 2F2, the wherein neuron number of this layer is 4096, by the way of stochastic gradient descent；
Softmax output layer O1, using Softmax return mode, neuron number be 4, i.e., with 4 graded category phases of American Ginseng
Corresponding, the value of output belongs to the probability value of 4 graded category American Ginseng images for American Ginseng sample image to be identified.
Claims (5)
- It is 1. a kind of based on wechat small routine and the American Ginseng of deep learning technology tasting assistant, it is characterised in that：Include installation There are the mobile equipment, American Ginseng tasting assistant's small routine terminal and cloud server analysis platform of wechat client；The shifting Dynamic equipment, using wechat client, it is whole to scan the two-dimensional code American Ginseng tasting assistant small routine described in load operating by network connection End, American Ginseng tasting assistant's small routine terminal call the camera of the mobile equipment to shoot American Ginseng sample to be identified Image, and the cloud server analysis platform is uploaded to, the cloud server analysis platform judges depth using American Ginseng Learning neural network model carries out American Ginseng sample image to be identified level identification, and qualification result is returned to the shifting Dynamic terminal, and show analysis result in American Ginseng tasting assistant's small routine terminal.
- It is 2. according to claim 1 based on wechat small routine and the American Ginseng of deep learning technology tasting assistant, feature It is：The interface of American Ginseng tasting assistant's small routine terminal is by shoot button, preview picture area and results display area group Into；Wherein, results display area can list 4 graded categories and two parameters of confidence level, and be arranged from high to low by confidence level.
- It is 3. according to claim 1 based on wechat small routine and the American Ginseng of deep learning technology tasting assistant, feature It is：The cloud server analysis platform includes image and receives and preprocessing module, deep neural network analysis module With result loopback module；It is the deep learning neural network analysis mould that described image, which is received with the output data of preprocessing module, The input data of block, wherein, described image, which is received, to be used to receiving the temporary mobile equipment upload with preprocessing module and waits to know Simultaneously preceding pretreatment is identified in other American Ginseng sample image, wherein the content pre-processed includes：It is sized, data channel is handed over It changes, the exchange of Color Channel；The result loopback module is used to extract point of the deep learning neural network analysis module Result data is analysed, and arranges and transfers to American Ginseng tasting assistant's small routine through wechat public platform server for JSON forms Terminal.
- It is 4. according to claim 1 based on wechat small routine and the American Ginseng of deep learning technology tasting assistant, feature It is：The preparation method such as following steps of the American Ginseng tasting deep learning neural network model：American Ginseng is divided into first, second and third graded category, and by above three by step 1 according to American Ginseng commercial specification index Graded category is collected and mark American Ginseng sample image, and collects American Ginseng Its Common Confused Varieties image, and Its Common Confused Varieties are marked For fourth estate classification, image data base is established, wherein, the image data of each graded category is not less than 1000 width；The pretreatment of step 2, image data sample, including：By image using mirror image rotation, scaling, lighting simulation to picture number Sample size amplification is realized according to library；256*256 sizes are uniformly scaled, and are converted into LMDB formatted datas；Step 3, establishment are upset by the label text file of " image path and class label " form and are ranked sequentially label text text " image path and class label " record strip in part, to the record strip of each graded category, by 5：1 ratio extracts training set With test set；Step 4, using Caffe deep learning frames, according to the parameter initialization pre-training neural network of AlexNet models, and The output layer neuron number of classification is changed to 4；Tuning is carried out, and protect to network parameter using the training set that step 3 obtains It deposits, export American Ginseng tasting deep learning neural network model；Step 5, beyond the clouds arrangement step 4 obtain American Ginseng tasting deep learning neural network model.
- 5. assistant is judged based on wechat small routine and the American Ginseng of deep learning technology according to claim 1 or 4, it is special Sign is：The American Ginseng tasting deep learning neural network model includes：Input layer, convolutional layer 1,2,3,4,5, pond layer 1st, 2,3, connect layer 1,2 and Softmax output layers entirely；The input layer, convolutional layer 1, pond layer 1, convolutional layer 2, pond layer 2, Convolutional layer 3, convolutional layer 4, convolutional layer 5, pond layer 3, it is complete connect layer 1, the full layer 2 that connects is connected with Softmax output layers sequence, upper level Output data be next stage input data；Each input picture of the input layer divides variability size for 227*227, Divide tri- color dimensions inputs of RGB；The convolutional layer 1, convolution nuclear volume are 96, and the size of each convolution kernel is 11*11, Convolution stride is 4；The pond layer 1, using maximization pond mode, Qi Chihua ranging from 3*3, sampling step length 2；It is described Convolutional layer 2, convolution nuclear volume is 256, and the size of each convolution kernel is 5*5, and convolution group is 2, and it is 2 to expand edge；It is described Pond layer 2, using maximize pond mode, Qi Chihua ranging from 3*3, sampling step length 2；The convolutional layer 3, convolution Nuclear volume is 384, and the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；The convolutional layer 4, convolution Nuclear volume is 384, and the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；The convolutional layer 5, convolution Nuclear volume is 256, and the size of each convolution kernel is 3*3, and convolution group is 2, and it is 2 to expand edge；The pond layer 3, using most Bigization pond mode, Qi Chihua ranging from 3*3, sampling step length 2；Described connects layer 1 entirely, and the neuron number of wherein this layer is 4096, by the way of stochastic gradient descent；Described connects layer 2 entirely, and the wherein neuron number of this layer is 4096, is used The mode of stochastic gradient descent；The Softmax output layers return mode using Softmax, and neuron number is 4, i.e., Corresponding with 4 graded categories of American Ginseng, the value of output belongs to 4 graded category west for American Ginseng sample image to be identified The probability value of American ginseng image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810066788.4A CN108229979A (en) | 2018-01-24 | 2018-01-24 | Based on wechat small routine and the American Ginseng of depth learning technology tasting assistant |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810066788.4A CN108229979A (en) | 2018-01-24 | 2018-01-24 | Based on wechat small routine and the American Ginseng of depth learning technology tasting assistant |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108229979A true CN108229979A (en) | 2018-06-29 |
Family
ID=62668710
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810066788.4A Pending CN108229979A (en) | 2018-01-24 | 2018-01-24 | Based on wechat small routine and the American Ginseng of depth learning technology tasting assistant |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108229979A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120783A (en) * | 2018-08-10 | 2019-01-01 | 深圳市小满科技有限公司 | Information acquisition method and device, mobile terminal and computer readable storage medium |
| CN110310135A (en) * | 2019-07-04 | 2019-10-08 | 山东浪潮人工智能研究院有限公司 | A kind of traceability system and method based on the study of potato picture depth |
| CN110532441A (en) * | 2019-08-23 | 2019-12-03 | 广州医科大学 | A kind of electronic component wisdom management method and its system |
| CN112530580A (en) * | 2020-12-03 | 2021-03-19 | 上海联影智能医疗科技有限公司 | Medical image picture processing method and computer readable storage medium |
| CN115587298A (en) * | 2021-07-05 | 2023-01-10 | 中国矿业大学(北京) | Historical Jingdezhen blue and white porcelain age discrimination method based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512676A (en) * | 2015-11-30 | 2016-04-20 | 华南理工大学 | Food recognition method at intelligent terminal |
| CN107153844A (en) * | 2017-05-12 | 2017-09-12 | 上海斐讯数据通信技术有限公司 | The accessory system being improved to flowers identifying system and the method being improved |
| CN107247957A (en) * | 2016-12-16 | 2017-10-13 | 广州中国科学院先进技术研究所 | A kind of intelligent agricultural product sorting technique and system based on deep learning and cloud computing |
| CN107590754A (en) * | 2017-11-01 | 2018-01-16 | 首都师范大学 | A kind of system and method based on augmented reality lifting national park Tourist Experience |
-
2018
- 2018-01-24 CN CN201810066788.4A patent/CN108229979A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512676A (en) * | 2015-11-30 | 2016-04-20 | 华南理工大学 | Food recognition method at intelligent terminal |
| CN107247957A (en) * | 2016-12-16 | 2017-10-13 | 广州中国科学院先进技术研究所 | A kind of intelligent agricultural product sorting technique and system based on deep learning and cloud computing |
| CN107153844A (en) * | 2017-05-12 | 2017-09-12 | 上海斐讯数据通信技术有限公司 | The accessory system being improved to flowers identifying system and the method being improved |
| CN107590754A (en) * | 2017-11-01 | 2018-01-16 | 首都师范大学 | A kind of system and method based on augmented reality lifting national park Tourist Experience |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120783A (en) * | 2018-08-10 | 2019-01-01 | 深圳市小满科技有限公司 | Information acquisition method and device, mobile terminal and computer readable storage medium |
| CN110310135A (en) * | 2019-07-04 | 2019-10-08 | 山东浪潮人工智能研究院有限公司 | A kind of traceability system and method based on the study of potato picture depth |
| CN110532441A (en) * | 2019-08-23 | 2019-12-03 | 广州医科大学 | A kind of electronic component wisdom management method and its system |
| CN112530580A (en) * | 2020-12-03 | 2021-03-19 | 上海联影智能医疗科技有限公司 | Medical image picture processing method and computer readable storage medium |
| CN115587298A (en) * | 2021-07-05 | 2023-01-10 | 中国矿业大学(北京) | Historical Jingdezhen blue and white porcelain age discrimination method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108229979A (en) | Based on wechat small routine and the American Ginseng of depth learning technology tasting assistant | |
| CN110070101B (en) | Plant species identification method and device, storage medium and computer equipment | |
| CN109033998B (en) | Remote sensing image ground object labeling method based on attention mechanism convolutional neural network | |
| CN105118048B (en) | The recognition methods of reproduction certificate picture and device | |
| CN109840560B (en) | Image classification method based on clustering in capsule network | |
| CN110967339B (en) | Method and device for analyzing corn ear characters and corn character analysis equipment | |
| CN111709816A (en) | Service recommendation method, device and equipment based on image recognition and storage medium | |
| CN109191428A (en) | Full-reference image quality evaluating method based on masking textural characteristics | |
| CN107274373B (en) | Code printing method and device in live streaming | |
| CN110647941A (en) | Vegetable identification method and equipment based on convolutional neural network | |
| CN109685713A (en) | Makeup analog control method, device, computer equipment and storage medium | |
| CN106650795A (en) | Sorting method of hotel room type images | |
| CN106445977A (en) | Picture pushing method and device | |
| CN106645180A (en) | Method for checking defects of substrate glass, field terminal and server | |
| CN104361357A (en) | Photo set classification system and method based on picture content analysis | |
| CN110059715A (en) | Floristic recognition methods and device, storage medium, computer equipment | |
| CN105825228B (en) | Image-recognizing method and device | |
| CN107424117A (en) | Image U.S. face method, apparatus, computer-readable recording medium and computer equipment | |
| CN108256557A (en) | The hyperspectral image classification method integrated with reference to deep learning and neighborhood | |
| CN109948676B (en) | Artificial intelligence-based traditional Chinese medicine decoction piece planting production area identification method | |
| CN111222545A (en) | Image classification method based on linear programming incremental learning | |
| CN106503747A (en) | A kind of image recognition statistical analysis system | |
| US20210357828A1 (en) | Method, server and system for providing feedback on user's mark on file | |
| CN110991434B (en) | Self-service terminal certificate identification method and device | |
| Bhatti et al. | Mobile cosmetics advisor: an imaging based mobile service |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180629 |
|
| RJ01 | Rejection of invention patent application after publication |