CN110363073A - Ligulate recognition methods, device, computer equipment and computer readable storage medium - Google Patents
Ligulate recognition methods, device, computer equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN110363073A CN110363073A CN201910468448.9A CN201910468448A CN110363073A CN 110363073 A CN110363073 A CN 110363073A CN 201910468448 A CN201910468448 A CN 201910468448A CN 110363073 A CN110363073 A CN 110363073A
- Authority
- CN
- China
- Prior art keywords
- ligulate
- layer
- convolutional layer
- convolutional
- tongue image
- 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
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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of ligulate recognition methods, comprising: obtains tongue image;Tongue image is pre-processed;According to the tongue image recognition ligulate, wherein, by the way that the tongue image is inputted ligulate identification model to obtain ligulate, the ligulate identification model is lightweight convolutional neural networks, and the lightweight convolutional neural networks use the semantic segmentation method based on depth convolutional neural networks.The invention also discloses a kind of ligulate identification device, computer equipment and computer readable storage mediums.Using the present invention, user can carry out ligulate identification whenever and wherever possible, and flow operations are simple, and ligulate identification identification is accurate.
Description
Technical field
The present invention relates to artificial intelligence ligulate identification fields more particularly to a kind of ligulate recognition methods, device, computer to set
Standby and computer readable storage medium.
Background technique
Lingual diagnosis is one of very important diagnostic techniques in Chinese medicine, wherein the diagnosis to ligulate is more particularly important, Ke Yitong
It crosses ligulate and quickly understands physical condition.User requires actively to go to look for a doctor and check if it is intended to carry out lingual diagnosis at present, and nothing
Method oneself diagnosis, is unfavorable for patient and understands itself physical condition in time, is also unfavorable for the state of an illness tracking of doctor follow-up patient.Simultaneously
Sub-health population is many in society, and many major diseases are all due to usually last without attention inferior health problem in fact
Major disease is caused.Traditional artificial lingual diagnosis mode is limited by geographical location, in addition artificial intelligence technology Shang Bufa in the past
It reaches, diagnosis identification directly can not be carried out by technological means.In addition, traditional lingual diagnosis instrument, it is big, inconvenient to carry, single that there are weight
People can not operate, service the problems such as single.
Summary of the invention
Technical problem to be solved by the present invention lies in provide ligulate recognition methods, device, computer equipment and computer
Readable storage medium storing program for executing, is able to solve that weight present in traditional lingual diagnosis instrument is big, inconvenient to carry, one can not operate, it is single to service
As soon as the problems such as, user can carry out ligulate identification by mobile terminals such as mobile phones whenever and wherever possible.
In order to solve the above-mentioned technical problems, the present invention provides a kind of ligulate recognition methods, comprising: obtains tongue figure
Picture;Tongue image is pre-processed;According to the tongue image recognition ligulate, wherein by inputting the tongue image
For ligulate identification model to obtain ligulate, the ligulate identification model is lightweight convolutional neural networks, the lightweight convolution mind
The semantic segmentation method based on depth convolutional neural networks is used through network.
Preferably, the network architecture model of the ligulate identification model includes sequentially connected first convolutional layer, the first pond
Change layer, the second convolutional layer, the second pond layer, third convolutional layer, third pond layer, Volume Four lamination, the first up-sampling layer, the 5th
Convolutional layer, the first articulamentum, the 6th convolutional layer, the second up-sampling layer, the 7th convolutional layer, the second articulamentum, the 8th convolutional layer, the
Three up-sampling layers, the 9th convolutional layer, third articulamentum, the tenth convolutional layer and the 11st convolutional layer, the ligulate identification model
It is trained using Adam optimization algorithm, and designed double loss functions is combined to export ligulate, in which: first up-sampling
Layer, the second up-sampling layer and third up-sampling layer are all made of bilinearity difference, and up-sampling multiplying power is 2;First convolution
Layer, the 9th convolutional layer and the tenth convolutional layer channel be 64, kernel be 3, stride be 1, the volume Two
It is 3, stride is 1 that the channel of lamination, the 7th convolutional layer and the 8th convolutional layer, which is 128, kernel, and described
It is 3, stride is 1 that the channel of three convolutional layers, the 5th convolutional layer and the 6th convolutional layer, which is 256, kernel, institute
The channel for stating Volume Four lamination is 512, kernel 3, stride 1, the channel of the 11st convolutional layer is 4,
Kernel is 1, stride 1, is connect with activation primitive after all convolutional layers;First pond layer, the second pond layer with
And the kernel and stride of third pond layer are 2;First articulamentum is by channel connection third convolutional layer and volume five
Lamination, second articulamentum connect the second convolutional layer and the 7th convolutional layer by channel, and the third articulamentum is connected by channel
First convolutional layer and the 9th convolutional layer.
Preferably, the learning rate of the Adam optimization algorithm is 0.0001, first order exponential matrix attenuation rate beta_1=
0.9, bi-exponential matrix attenuation rate beta_2=0.99, total 20 epochs of iteration.
Preferably, double loss functions include cross entropy loss function and IOU loss function, in which: the cross entropy
Loss function are as follows: fcross=-sum (y*log (p)), wherein when y is true value, y=1, otherwise y=0;The IOU loss
Function are as follows: fiou=1-2* (intersection+smooth)/(union+smooth), wherein intersection is prediction
Pixel region and real estate intersection, union be prediction pixel region and real estate union, smooth is smooth
The factor, value 1.
Preferably, the cross entropy loss function and IOU loss function are weighted by the coefficient of 0.9 and 0.1, i.e., finally
Loss function is f=0.9*fcross+0.1*fiou.
Preferably, described includes: by tongue image normalization to 256X256 size to tongue image progress pre-treatment step
Size, and pixel value is normalized into 0~1 section.
Correspondingly, the present invention also provides a kind of ligulate identification devices, comprising: module is obtained, for obtaining tongue image;
Preprocessing module, for tongue image normalization to be normalized to 0th~1 area to 256X256 size, and by pixel value
Between;Identification module, for according to the defeated identification ligulate of the tongue image, wherein known by the way that the tongue image is inputted ligulate
For other model to obtain ligulate, the ligulate identification model is lightweight convolutional neural networks, the lightweight convolutional neural networks
Using the semantic segmentation method based on depth convolutional neural networks.
Preferably, the ligulate identification model in the ligulate identification module includes sequentially connected first convolutional layer, first
Pond layer, the second convolutional layer, the second pond layer, third convolutional layer, third pond layer, Volume Four lamination, the first up-sampling layer, the
Five convolutional layers, the first articulamentum, the 6th convolutional layer, second up-sampling layer, the 7th convolutional layer, the second articulamentum, the 8th convolutional layer,
Third up-samples layer, the 9th convolutional layer, third articulamentum, the tenth convolutional layer and the 11st convolutional layer.
Correspondingly, the present invention also provides a kind of computer equipment, including memory and processor, the memory storages
There is computer program, which is characterized in that the processor is realized any in claim 1 to 8 when executing the computer program
The step of method described in item.
Correspondingly, the present invention also provides a kind of computer readable storage mediums, are stored thereon with computer program, special
The step of sign is, the computer program realizes method described in any item of the claim 1 to 8 when being executed by processor.
The beneficial effects of the practice of the present invention is:
The present invention is by artificial intelligence image recognition technology, by the way that depth convolutional neural networks model is introduced ligulate identification
Field, user only need to clap a tongue photo by mobile terminal, so that it may know ligulate recognition result, solve traditional tongue diagnosing instrument
Device carries the problems such as difficulty, and user only need to be in corresponding webpage or app, so that it may ligulate identification is carried out whenever and wherever possible, it is easy to operate, and know
It is inaccurate, comprehensive health service is provided for user.
Detailed description of the invention
Fig. 1 is ligulate recognition methods flow diagram provided by the invention;
Fig. 2 is ligulate identification model schematic diagram provided by the invention;
Fig. 3 is ligulate identification device schematic diagram provided by the invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, the present invention is made into one below in conjunction with attached drawing
Step ground detailed description.Only this is stated, the present invention occurs in the text or will occur up, down, left, right, before and after, it is inside and outside etc. just
Position word is not to specific restriction of the invention only on the basis of attached drawing of the invention.
As shown in Figure 1, the present invention provides a kind of ligulate recognition methods, comprising:
S101 obtains tongue image.
It should be noted that tongue image can be obtained in real time by the camera on mobile terminal (mobile phone terminal or the end PC) or
Tongue image is normalized after obtaining tongue image by local upload.
Specifically, when obtaining tongue image in real time by camera, what user need to only be put into tongue that mobile terminal shows is obtained
In contouring, wherein in order to preferably obtain tongue image, obtain profile and preferentially use class parabolic profile.
S102 pre-processes tongue image.
After obtaining tongue image, tongue image normalization need to be handled, the present invention preferentially selects picture size to normalize to
256X256, pixel value normalize to 0~1 section.
S103, according to the tongue image recognition ligulate.
It should be noted that after normalized, by the way that the tongue image is inputted ligulate identification model to obtain tongue
Shape, the ligulate identification model are lightweight convolutional neural networks, and the lightweight convolutional neural networks are used to be rolled up based on depth
The semantic segmentation method of product neural network.
Specifically, the ligulate identification model is trained using Adam optimization algorithm, and combine designed double losses
Function exports ligulate, in which:
The learning rate of the Adam optimization algorithm is 0.0001, and first order exponential matrix attenuation rate beta_1=0.9, second order refers to
Matrix number attenuation rate beta_2=0.99, total 20 epochs of iteration;
Double loss functions include cross entropy loss function and IOU loss function, in which: the cross entropy loss function
Are as follows: fcross=-sum (y*log (p)), wherein when y is true value, y=1, otherwise y=0;The IOU loss function are as follows:
Fiou=1-2* (intersection+smooth)/(union+smooth), wherein intersection is the pixel of prediction
The intersection in region and real estate, union are the pixel region of prediction and the union of real estate, and smooth is smoothing factor,
Value is 1;The cross entropy loss function and IOU loss function are weighted by the coefficient of 0.9 and 0.1, i.e., finally lose letter
Number is f=0.9*fcross+0.1*fiou.
As shown in Fig. 2, the network architecture model of the ligulate identification model includes sequentially connected first convolutional layer, first
Pond layer, the second convolutional layer, the second pond layer, third convolutional layer, third pond layer, Volume Four lamination, the first up-sampling layer, the
Five convolutional layers, the first articulamentum, the 6th convolutional layer, second up-sampling layer, the 7th convolutional layer, the second articulamentum, the 8th convolutional layer,
Third up-samples layer, the 9th convolutional layer, third articulamentum, the tenth convolutional layer and the 11st convolutional layer, in which: on described first
Sample level, the second up-sampling layer and third up-sampling layer are all made of bilinearity difference, and up-sampling multiplying power is 2;Described first
It is 3, stride is 1 that the channel of convolutional layer, the 9th convolutional layer and the tenth convolutional layer, which is 64, kernel, and described
It is 3, stride is 1 that the channel of two convolutional layers, the 7th convolutional layer and the 8th convolutional layer, which is 128, kernel, institute
It is that 3, stride is that the channel for stating third convolutional layer, the 5th convolutional layer and the 6th convolutional layer, which is 256, kernel,
1, the channel of the Volume Four lamination be 512, kernel 3, stride 1, the channel of the 11st convolutional layer
For 4, kernel 1, stride 1, it is connect with activation primitive after all convolutional layers.First pond layer, the second pond
The kernel and stride of layer and third pond layer are 2;First articulamentum is by channel connection third convolutional layer and the
Five convolutional layers, second articulamentum connect the second convolutional layer and the 7th convolutional layer by channel, and the third articulamentum presses channel
Connect the first convolutional layer and the 9th convolutional layer.
It should be noted that ligulate identification process are as follows: after the tongue image normalization that will test out, be sent into designed volume
In product neural network model, ligulate characteristic pattern is exported, by the threshold value manually set, binaryzation is carried out to ligulate characteristic pattern, i.e.,
Exportable ligulate location information.Profile is extracted based on this bianry image, so as to calculate the mathematical feature of ligulate (such as crackle
The information such as quantity, length, the quantity of spot, indentation).
As shown in figure 3, the present invention provides a kind of ligulate identification devices 100, comprising:
Obtain module 1, for obtaining tongue image, it should be noted that tongue image can by mobile terminal (mobile phone terminal or
The end PC) on camera obtain or uploaded by local in real time, after obtaining tongue image, place is normalized to tongue image
Reason.Specifically, when obtaining tongue image in real time by camera, user need to only be put into tongue the acquisition profile that mobile terminal is shown
In, wherein in order to preferably obtain tongue image, obtains profile and preferentially use class parabolic profile.
Preprocessing module 2, for normalizing tongue image normalization to 256X256 size, and by pixel value
To 0~1 section.
Ligulate identification module 3, for according to the defeated identification ligulate of the tongue image, wherein by by the tongue image
For input ligulate identification model to obtain ligulate, the ligulate identification model is lightweight convolutional neural networks, the lightweight volume
Product neural network uses the semantic segmentation method based on depth convolutional neural networks.Specifically, in the ligulate identification module
Ligulate identification model includes sequentially connected first convolutional layer, the first pond layer, the second convolutional layer, the second pond layer, third volume
Lamination, third pond layer, Volume Four lamination, the first up-sampling layer, the 5th convolutional layer, the first articulamentum, the 6th convolutional layer, second
Up-sample layer, the 7th convolutional layer, the second articulamentum, the 8th convolutional layer, third up-sampling layer, the 9th convolutional layer, third articulamentum,
Tenth convolutional layer and the 11st convolutional layer.The ligulate identification model is trained using Adam optimization algorithm, and is combined and set
The double loss functions output ligulate counted.
Correspondingly, the present invention also provides a kind of computer equipment, including memory and processor, the memory storage
There is the step of computer program, the processor realizes above-mentioned ligulate recognition methods when executing the computer program.Meanwhile this
Invention additionally provides a kind of computer readable storage medium, is stored thereon with computer program, the computer program is processed
The step of device realizes above-mentioned ligulate recognition methods when executing.
From the foregoing, it will be observed that the present invention can be used for the identifying processing to ligulate, the present invention passes through artificial intelligence image recognition technology,
Field is identified by the way that depth convolutional neural networks model is introduced ligulate, user only needs to clap a tongue photo by mobile terminal,
It is known that ligulate recognition result, solve traditional lingual diagnosis instrument and carry the problems such as difficulty, user only need in corresponding webpage or
App, so that it may carry out ligulate identification whenever and wherever possible, easy to operate, identification is accurate, and comprehensive health service is provided for user.
The above is a preferred embodiment of the present invention, it is noted that for those skilled in the art
For, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (10)
1. a kind of ligulate recognition methods characterized by comprising
Obtain tongue image;
Tongue image is pre-processed;
According to the tongue image recognition ligulate, wherein by the way that the tongue image is inputted ligulate identification model to obtain tongue
Shape, the ligulate identification model are lightweight convolutional neural networks, and the lightweight convolutional neural networks are used to be rolled up based on depth
The semantic segmentation method of product neural network.
2. ligulate recognition methods as described in claim 1, which is characterized in that the network architecture model of the ligulate identification model
Including sequentially connected first convolutional layer, the first pond layer, the second convolutional layer, the second pond layer, third convolutional layer, third pond
Layer, Volume Four lamination, the first up-sampling layer, the 5th convolutional layer, the first articulamentum, the 6th convolutional layer, the second up-sampling layer, the 7th
Convolutional layer, the second articulamentum, the 8th convolutional layer, third up-sampling layer, the 9th convolutional layer, third articulamentum, the tenth convolutional layer with
And the 11st convolutional layer, the ligulate identification model are trained using Adam optimization algorithm, and combine designed double losses
Function exports ligulate, in which:
The first up-sampling layer, the second up-sampling layer and third up-sampling layer are all made of bilinearity difference, up-sample multiplying power
It is 2;
It is 3, stride that the channel of first convolutional layer, the 9th convolutional layer and the tenth convolutional layer, which is 64, kernel,
Be 1, the channel of second convolutional layer, the 7th convolutional layer and the 8th convolutional layer be 128, kernel be 3,
Stride is 1, and the channel of the third convolutional layer, the 5th convolutional layer and the 6th convolutional layer is 256, kernel equal
It is 1 for 3, stride, the channel of the Volume Four lamination is 512, kernel 3, stride 1, and the described tenth is a roll of
The channel of lamination is 4, kernel 1, stride 1, is connect with activation primitive after all convolutional layers;
The kernel and stride of first pond layer, the second pond layer and third pond layer are 2;
First articulamentum is by channel connection third convolutional layer and the 5th convolutional layer, and second articulamentum is by channel connection the
Two convolutional layers and the 7th convolutional layer, the third articulamentum connect the first convolutional layer and the 9th convolutional layer by channel.
3. ligulate recognition methods as claimed in claim 2, which is characterized in that the learning rate of the Adam optimization algorithm is
0.0001, first order exponential matrix attenuation rate beta_1=0.9, bi-exponential matrix attenuation rate beta_2=0.99, total iteration 20
A epochs.
4. ligulate recognition methods as claimed in claim 3, which is characterized in that double loss functions include intersecting entropy loss letter
Several and IOU loss function, in which:
The cross entropy loss function are as follows:
Fcross=-sum (y*log (p)),
Wherein, when y is true value, y=1, otherwise y=0;
The IOU loss function are as follows:
Fiou=1-2* (intersection+smooth)/(union+smooth),
Wherein, intersection be prediction pixel region and real estate intersection, union be prediction pixel region and
The union of real estate, smooth are smoothing factor, value 1.
5. ligulate recognition methods as claimed in claim 4, which is characterized in that the cross entropy loss function and IOU lose letter
Number is weighted by the coefficient of 0.9 and 0.1, i.e., final loss function is f=0.9*fcross+0.1*fiou.
6. ligulate recognition methods as described in claim 1, which is characterized in that described to carry out pre-treatment step packet to tongue image
It includes: by tongue image normalization to 256X256 size, and pixel value being normalized into 0~1 section.
7. a kind of ligulate identification device characterized by comprising
Module is obtained, for obtaining tongue image;
Preprocessing module, for tongue image normalization to be normalized to 0~1 to 256X256 size, and by pixel value
Section;
Identification module, for according to the defeated identification ligulate of the tongue image, wherein known by the way that the tongue image is inputted ligulate
For other model to obtain ligulate, the ligulate identification model is lightweight convolutional neural networks, the lightweight convolutional neural networks
Using the semantic segmentation method based on depth convolutional neural networks.
8. ligulate identification device as claimed in claim 7, which is characterized in that the ligulate in the ligulate identification module identifies mould
Type includes sequentially connected first convolutional layer, the first pond layer, the second convolutional layer, the second pond layer, third convolutional layer, third pond
Change layer, Volume Four lamination, the first up-sampling layer, the 5th convolutional layer, the first articulamentum, the 6th convolutional layer, the second up-sampling layer, the
Seven convolutional layers, the second articulamentum, the 8th convolutional layer, third up-sample layer, the 9th convolutional layer, third articulamentum, the tenth convolutional layer
And the 11st convolutional layer.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes method described in any item of the claim 1 to 8 when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any item of the claim 1 to 8 is realized when being executed by processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910468448.9A CN110363073B (en) | 2019-05-31 | 2019-05-31 | Tongue recognition method, tongue recognition device, computer equipment and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910468448.9A CN110363073B (en) | 2019-05-31 | 2019-05-31 | Tongue recognition method, tongue recognition device, computer equipment and computer readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110363073A true CN110363073A (en) | 2019-10-22 |
CN110363073B CN110363073B (en) | 2023-04-25 |
Family
ID=68215439
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910468448.9A Active CN110363073B (en) | 2019-05-31 | 2019-05-31 | Tongue recognition method, tongue recognition device, computer equipment and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110363073B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113744276A (en) * | 2020-05-13 | 2021-12-03 | Oppo广东移动通信有限公司 | Image processing method, image processing apparatus, electronic device, and readable storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105286768A (en) * | 2015-09-23 | 2016-02-03 | 江苏大学 | Human health status tongue coating diagnosis device based on mobile phone platform |
CN109635711A (en) * | 2018-12-07 | 2019-04-16 | 上海衡道医学病理诊断中心有限公司 | A kind of pathological image dividing method based on deep learning network |
-
2019
- 2019-05-31 CN CN201910468448.9A patent/CN110363073B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105286768A (en) * | 2015-09-23 | 2016-02-03 | 江苏大学 | Human health status tongue coating diagnosis device based on mobile phone platform |
CN109635711A (en) * | 2018-12-07 | 2019-04-16 | 上海衡道医学病理诊断中心有限公司 | A kind of pathological image dividing method based on deep learning network |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113744276A (en) * | 2020-05-13 | 2021-12-03 | Oppo广东移动通信有限公司 | Image processing method, image processing apparatus, electronic device, and readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN110363073B (en) | 2023-04-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11694123B2 (en) | Computer based object detection within a video or image | |
WO2022105714A1 (en) | Data processing method, machine learning training method and related apparatus, and device | |
CN112329964B (en) | Method, device, equipment and storage medium for pushing information | |
WO2021218471A1 (en) | Neural network for image processing and related device | |
CN110363072A (en) | Tongue image recognition method, apparatus, computer equipment and computer readable storage medium | |
CN110660484A (en) | Bone age prediction method, device, medium, and electronic apparatus | |
CN112419326B (en) | Image segmentation data processing method, device, equipment and storage medium | |
CN114512228A (en) | Traditional Chinese medicine disease auxiliary diagnosis system, equipment and storage medium | |
WO2022111387A1 (en) | Data processing method and related apparatus | |
Meng et al. | Few-shot image classification algorithm based on attention mechanism and weight fusion | |
CN110110724A (en) | The text authentication code recognition methods of function drive capsule neural network is squeezed based on exponential type | |
CN114886404B (en) | Electronic equipment, device and storage medium | |
CN113656563A (en) | Neural network searching method and related equipment | |
AU2024204390A1 (en) | Computer based object detection within a video or image | |
CN115775349A (en) | False news detection method and device based on multi-mode fusion | |
CN111160049A (en) | Text translation method, device, machine translation system and storage medium | |
CN110363073A (en) | Ligulate recognition methods, device, computer equipment and computer readable storage medium | |
CN116913459B (en) | Medicine recommendation method and system based on deep convolution network control gate model | |
CN116469534A (en) | Hospital number calling management system and method thereof | |
CN115659987A (en) | Multi-mode named entity recognition method, device and equipment based on double channels | |
Yu et al. | Pyramidal and conditional convolution attention network for hyperspectral image classification using limited training samples | |
CN116994695A (en) | Training method, device, equipment and storage medium of report generation model | |
CN116955635A (en) | Text matching method, device, equipment, storage medium and program product | |
CN118132727A (en) | Dialogue data processing method, device, equipment and medium | |
CN116091510A (en) | Retina OCT image layering model training method, retina OCT image layering model training device, computer equipment and 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 | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20230407 Address after: No. 53 B2, No. 397, Xingang Middle Road, Haizhu District, Guangzhou City, Guangdong Province (office only) Applicant after: Guangdong and state network technology Co.,Ltd. Address before: Room 2801, No. 28-1, Machang Road, Tianhe District, Guangzhou, Guangdong 510000 Applicant before: Zhenghe intelligent network technology (Guangzhou) Co.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |