CN109712707A - A kind of lingual diagnosis method, apparatus calculates equipment and computer storage medium - Google Patents
A kind of lingual diagnosis method, apparatus calculates equipment and computer storage medium Download PDFInfo
- Publication number
- CN109712707A CN109712707A CN201811634491.XA CN201811634491A CN109712707A CN 109712707 A CN109712707 A CN 109712707A CN 201811634491 A CN201811634491 A CN 201811634491A CN 109712707 A CN109712707 A CN 109712707A
- Authority
- CN
- China
- Prior art keywords
- tongue
- network
- image
- positioning
- neural network
- 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
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The present invention relates to field of artificial intelligence, in particular discloses a kind of lingual diagnosis method, apparatus, calculates equipment and computer storage medium, wherein method includes: the tongue image for obtaining user；The tongue image is inputted into the positioning tongue image after positioning neural network obtains the positioning of tongue target, the positioning neural network is obtained by first nerves network training, and the first nerves network includes SSD neural network；The positioning tongue image is inputted into identification neural network and obtains recognition result, the identification neural network is obtained by nervus opticus network training, the first nerves network and the nervus opticus network are same neural network or the first nerves network and the nervus opticus network is different neural networks.It can be seen that wisdom lingual diagnosis may be implemented using the present invention program.
Description
Technical field
The present embodiments relate to field of artificial intelligence, more particularly to a kind of lingual diagnosis method, apparatus, calculate equipment
And computer storage medium.
Background technique
Lingual diagnosis is the color for observing tongue, the variation of form carrys out auxiliary diagnosis and one of identifying disease simply and effectively square
Method is one of method for tcm diagnosis.In recent years, Chinese medicine intelligent diagnosis system came out, and was examined with intellectual technology auxiliary Chinese medicine
It is disconnected, to mitigate Chinese medicine hard work amount.Currently, lingual diagnosis technology is broadly divided into two parts: being based on traditional images location technology
Lingual diagnosis technology and lingual diagnosis technology instrument-based.
The present inventor in the implementation of the present invention, has found: the lingual diagnosis side based on traditional images location technology
Method generally uses Haar, and the traditional images location algorithm such as Snake, robustness is bad, is encountering different light environments or tongue shape
The case where shape and color change, cannot position well；Instrument price is expensive, needs to shoot patient under fixed environment within the hospital
Tongue image, effect is good, but user is inconvenient to use.
Summary of the invention
In view of the above problems, it proposes on the present invention overcomes the above problem or at least be partially solved in order to provide one kind
It states a kind of lingual diagnosis method, apparatus of problem, calculate equipment and computer storage medium.
In order to solve the above technical problems, a technical solution used in the embodiment of the present invention is: a kind of lingual diagnosis method is provided,
It include: the tongue image for obtaining user；The tongue image is inputted positioning neural network to obtain determining after tongue target positions
Position tongue image, the positioning neural network is obtained by first nerves network training, and the first nerves network includes
SSD neural network；The positioning tongue image is inputted into identification neural network and obtains recognition result, the identification neural network is
It being obtained by nervus opticus network training, the first nerves network and the nervus opticus network are same neural network,
Or the first nerves network is different neural networks from the nervus opticus network.
Optionally, the method also includes including: the tongue sample image for obtaining preset quantity；Mark the tongue sample
The circumscribed frame at tongue position in image obtains positioning tongue sample image；By the tongue sample image and the positioning tongue
Sample image inputs the first nerves network and is trained, and obtains the positioning neural network.
Optionally, the method also includes: obtain the positioning tongue sample image；To the positioning tongue sample image
Tongue label is added, the tongue label is used to indicate the physical condition of user corresponding with the tongue sample image；It will be described
Positioning tongue sample image and the tongue label corresponding with the positioning tongue sample image are input to the nervus opticus
Network is trained to obtain identification neural network.
Optionally, the method also includes: according to Haar algorithm, detection is located at the coverage of mobile terminal taking lens
Interior image whether include user tongue image；If comprising obtaining the tongue image of user.
Optionally, the method also includes: if the image in the coverage of the mobile terminal taking lens does not include
The tongue image of user prompts user to adjust the position of mobile terminal taking lens, until the mobile terminal taking lens
Image in coverage includes the tongue image of user.
Optionally, the method also includes: image cleaning is carried out to the tongue image and the tongue sample image, is returned
One change, dimensionality reduction and/or whitening processing.
In order to solve the above technical problems, another technical solution used in the embodiment of the present invention is: providing a kind of lingual diagnosis dress
It sets characterized by comprising obtain module: for obtaining the tongue image of user；Locating module: it is used for the tongue figure
The positioning tongue image after tongue target positions is obtained as input positions neural network, the positioning neural network is by first
What neural metwork training obtained, the first nerves network includes SSD neural network；Identification module: it is used for the positioning tongue
Head image input identification neural network obtains recognition result, and the identification neural network is obtained by nervus opticus network training
, the first nerves network and the nervus opticus network are same neural network or the first nerves network and institute
Stating nervus opticus network is different neural networks.
Optionally, positioning neural network described in the locating module is obtained by first nerves network training, packet
It includes: the positioning neural network is obtained to the first nerves network training；It is described that the first nerves network training is obtained
The positioning neural network, comprising: obtain the tongue sample image of preset quantity；Mark tongue portion in the tongue sample image
The circumscribed frame of position obtains positioning tongue sample image；The tongue sample image and the positioning tongue sample image are inputted
The default first nerves network, obtains the positioning neural network.
Optionally, described in the identification module identification neural network be by being obtained to nervus opticus network training,
It include: that the identification neural network is obtained to the nervus opticus network training；It is described that the nervus opticus network training is obtained
To the identification neural network, comprising: obtain the positioning tongue sample image；Tongue is added to the positioning tongue sample image
Leader label, the tongue label are used to indicate the physical condition of user corresponding with the tongue sample image；By the positioning tongue
Head sample image and the tongue label corresponding with the positioning tongue sample image be input to the nervus opticus network into
Row training obtains identification neural network.
Optionally, described device further include: detection module: for according to Haar algorithm, detection to be located at mobile terminal and shoots
Image in the coverage of camera lens whether include user tongue image；Cue module: for being shot when the mobile terminal
Image in the coverage of camera lens does not include the tongue image of user, and user is prompted to adjust the position of mobile terminal taking lens
It sets, until the image in the coverage of the mobile terminal taking lens includes the tongue image of user.
Optionally, described device further include: image processing module: for the tongue image and the tongue sample graph
As carrying out image cleaning, normalization, dimensionality reduction and/or whitening processing.
In order to solve the above technical problems, another technical solution used in the embodiment of the present invention is: providing a kind of calculate and set
It is standby, comprising: processor, memory, communication interface and communication bus, the processor, the memory and the communication interface
Mutual communication is completed by the communication bus；
The memory makes described in the processor execution for storing an at least executable instruction, the executable instruction
A kind of corresponding operation of lingual diagnosis method.
In order to solve the above technical problems, another technical solution used in the embodiment of the present invention is: providing a kind of computer
Storage medium, is stored with an at least executable instruction in the storage medium, and the executable instruction makes described in processor executes
A kind of corresponding operation of lingual diagnosis method.
The beneficial effect of the embodiment of the present invention is: being in contrast to the prior art, the embodiment of the present invention passes through positioning mind
Physical condition through network and identification neural network recognization user, may be implemented wisdom lingual diagnosis, with existing some use nerves
The wisdom lingual diagnosis technology of network is compared, and the embodiment of the present invention is positioned using tongue image of the SSD neural network to user, is led to
It crosses and is identified comprising tongue image of the neural network including SSD neural network to user, obtain the physical condition of user, know
Other result is relatively reliable；In addition, using Haar algorithm detection customer mobile terminal taking lens coverage in whether contain
The tongue image of user can improve positioning nerve before positioning neural network positioning to the tongue image Primary Location of user
The efficiency of network, and the embodiment of the present invention be can be used for into mobile terminal, it is user-friendly.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention,
And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can
It is clearer and more comprehensible, it is special below to lift specific embodiments of the present invention.
Detailed description of the invention
By reading hereafter detailed description of preferred embodiment, various other advantages and benefits skill common for this field
Art personnel will become clear.Attached drawing is only used for showing the purpose of preferred embodiment, and is not considered as to limit of the invention
System.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is a kind of lingual diagnosis method flow diagram of the embodiment of the present invention；
Fig. 2 is that neural metwork training flow chart is positioned in a kind of lingual diagnosis method of the embodiment of the present invention；
Fig. 3 is that neural metwork training flow chart is identified in a kind of lingual diagnosis method of the embodiment of the present invention；
Fig. 4 is a kind of lingual diagnosis method flow diagram of another embodiment of the present invention；
Fig. 5 is a kind of lingual diagnosis apparatus function block diagram of the embodiment of the present invention；
Fig. 6 is a kind of calculating equipment schematic diagram of the embodiment of the present invention.
Specific embodiment
Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
It is fully disclosed to those skilled in the art.
Fig. 1 is a kind of lingual diagnosis method flow diagram of the embodiment of the present invention.As shown in Figure 1, method includes the following steps:
Step S101: the tongue image of user is obtained.
In this step, the tongue image of the user refers to the image comprising user's tongue position.
Step S102: the tongue image is inputted into the positioning tongue figure after positioning neural network obtains the positioning of tongue target
Picture.
In this step, the positioning neural network is obtained by first nerves network training, the first nerves
Network includes SSD neural network.
Fig. 2 shows neural metwork training flow chart is positioned in the embodiment of the present invention, as shown in Fig. 2, positioning neural network
Training the following steps are included:
Step S1021: the tongue sample image of preset quantity is obtained.
In this step, the tongue sample image is a large amount of tongues that research staff collected from hospital internal or on the net
Head sample image.
Step S1022: marking the circumscribed frame at tongue position in the tongue sample image, obtains positioning tongue sample graph
Picture.
In this step, tongue position in the tongue sample image is labeled using circumscribed frame, by the use
The tongue position image of circumscribed collimation mark note is as positioning tongue sample image.
Step S1023: the tongue sample image and the positioning tongue sample image are inputted into the first nerves net
Network is trained, and obtains the positioning neural network.
In this step, the default first nerves network can use any particular algorithms under SSD neural network framework,
Such as, mobileNet, ResNet, SquezzNet, ShuffleNet etc., by taking mobileNet as an example, when being trained, first
The training parameter in the mobileNet is defined, the training parameter includes anchor, sacle, ratio, wherein is being arranged
When anchor, according to the shape of tongue, rectangle is set by anchor box, e.g., a rectangle is indicated using Rec, makes
Indicate the rectangular length with x, y indicates the rectangular width, when being configured, guarantee x in the anchor box >
=y.Different scale and ratio is arranged in optimal anchor box in order to obtain, to obtain a series of
Anchor box, wherein the scale indicates the ratio of the long x Yu width y, ratio defines the scaling of the anchorbox
Ratio sets 32,16,8,4 four values for ratio in the present embodiment.Patient's tongue image is cut into fixed big
Small grid, wherein patient's tongue image is known as feature map, and the grid of the fixed size is known as feature
Map cell, and anchor box is exactly a series of box of fixed sizes on each feature map cell, by described one
Each of the box of serial fixed size box and the positioning tongue sample image, i.e. ground truth box comparison,
Obtain only one anchor box corresponding with the ground truth box, wherein the control methods is to calculate
IOU, the IOU refer to the degree of overlapping of each box Yu the ground truth box, best using calculated result as screening
The standard of box guarantees that each ground truth box is corresponding with only one anchor box；In the embodiment of the present invention
In, be arranged IOU threshold value be 0.85, i.e., when the IOU be greater than or equal to 0.85 when, determine the anchor box with it is described
Ground truth box is corresponding, wherein the anchor box is positive sample, remaining anchor box is negative sample.It determines
After anchorbox, the weight of the default first nerves network is calculated using stochastic gradient descent method, and use back-propagation
Algorithm updates the weight of each layer network in the default first nerves network, obtains positioning neural network.
Step S103: the positioning tongue image is inputted into identification neural network and obtains recognition result.
In this step, the identification neural network is obtained by nervus opticus network training, the first nerves
Network and the nervus opticus network are that same neural network or the first nerves network are with the nervus opticus network
Different neural networks, e.g., the nervus opticus network can be SSD neural network identical with the first nerves network,
It can be the neural network of some other mainstream, e.g., convolutional neural networks, Inception neural network etc..
Fig. 3 shows identification neural metwork training flow chart in the embodiment of the present invention, as shown in figure 3, identification neural network
Training the following steps are included:
Step S1031: the positioning tongue sample image is obtained.
In this step, the positioning tongue sample image is the tongue sample image to positioning neural metwork training
The circumscribed frame at middle tongue position obtains positioning tongue sample image.
Step S1032: tongue label is added to the positioning tongue sample image, the tongue label is used to indicate and institute
State the physical condition that tongue sample image corresponds to user.
In this step, the corresponding disease type of the positioning tongue sample image is known, by the positioning tongue
The corresponding physical condition type of sample image carries out label.In embodiments of the present invention, the physical condition type includes five big
Class: each when being labeled without disease, respiratory disease, disease of digestive system, disease in the urological system and heart disease
Kind disease corresponds to a kind of disease type label, 1 will be such as labeled as without disease, respiratory disease is labeled as 2, disease of digestive system
3 are labeled as, disease in the urological system is labeled as 4, and heart disease is labeled as 5.
Step S1033: by the positioning tongue sample image and the tongue corresponding with the positioning tongue sample image
Leader label are input to the nervus opticus network and are trained to obtain identification neural network.
In this step, when the nervus opticus network is identical as the first nerves network, by the positioning tongue
Sample image and the label corresponding with the positioning tongue sample image are input to the first nerves network and are trained,
It should be noted that if the nervus opticus network is identical as the first nerves network, then positioning neural network instruction is being carried out
When practicing, a large amount of sample image is needed, when the sample image quantity is very big, using first nerves network, including SSD nerve
Network is that can reach the purpose of medical diagnosis on disease, under this methodology, calculates the corresponding disease of each described positioning tongue sample image
The score of sick type thinks the positioning tongue sample image and the body shape when the score is higher than given threshold
Condition is corresponding.It when the positioning tongue sample image is less, is identified using nervus opticus network, specific training process is public affairs
Know technology, details are not described herein.In embodiments of the present invention, a large amount of sample had been collected, therefore, using SSD neural network
It is identified, the score threshold value is set as 0.8.
The embodiment of the present invention may be implemented by the physical condition of positioning neural network and identification neural network recognization user
Wisdom lingual diagnosis, compared with the more existing wisdom lingual diagnosis technology using neural network, the embodiment of the present invention uses SSD nerve net
Network positions the tongue image of user, by the inclusion of the neural network including SSD neural network to the tongue image of user
It is identified, obtain the physical condition of user, recognition result is relatively reliable.
Fig. 4 is a kind of lingual diagnosis method flow diagram of another embodiment of the present invention.As shown in figure 4, the embodiment and upper one is implemented
Example is compared, further comprising the steps of before the tongue image for obtaining user:
Step S401: according to Haar algorithm, whether detection is located at the image in the coverage of mobile terminal taking lens
Tongue image comprising user, if comprising executing step S402, if not including, executing step S403.
In this step, user shoots image using mobile terminal, and Haar algorithm is arranged in the mobile terminal, and real-time monitoring is used
It whether include tongue in the taking lens of family.When user shoots image, in order to obtain best shooting effect, tongue position is at whole
Ratio in image is 20%-40%, when detecting tongue proportion in whole image, user is prompted to remain stationary,
Obtain shooting image.
Step S402: the tongue image of user is obtained.
In this step, the tongue image of the user is that user shoots the tongue figure containing user's tongue position in image
Picture.
In one embodiment, image cleaning, normalization, dimensionality reduction are carried out to the tongue image and sample image of the user
And/or whitening processing, wherein described image cleaning is too low in order to not contain the image at tongue position, tongue image pixel
The too small image-erasing of image and tongue position accounting not easy to identify, meanwhile, in order to obtain more accurately as a result, when a figure
It, also can be corresponding image-erasing when containing multiple tongue positions as in；Normalized purpose is in order to by the tongue image
Zoom to identical size, is trained convenient for network.Dimensionality reduction and/or whitening processing are the complexity in order to reduce model training
Degree.
Step S403: prompt user adjusts the position of mobile terminal taking lens, until the mobile terminal taking lens
Coverage in image include user tongue image.
Whether the embodiment of the present invention contains in the coverage using Haar algorithm detection customer mobile terminal taking lens
The tongue image of user, can be preliminary to the tongue image of user before positioning neural network is to the tongue framing of user
Positioning, improves the efficiency of positioning neural network, and the embodiment of the present invention is used for mobile terminal, user-friendly.
Fig. 5 is a kind of lingual diagnosis apparatus function block diagram of the embodiment of the present invention, as shown in figure 4, described device includes: acquisition mould
Block 501, locating module 502, identification module 503, detection module 504, cue module 505 and image processing module 506, wherein
The acquisition module 501, for obtaining the tongue image of user；Locating module 502 is positioned for inputting the tongue image
Neural network obtains the positioning tongue image after the positioning of tongue target, and the positioning neural network is instructed by first nerves network
It gets, the first nerves network includes SSD neural network；Identification module 503, for the positioning tongue image is defeated
Enter to identify that neural network obtains recognition result, the identification neural network is obtained by nervus opticus network training, described
First nerves network and the nervus opticus network are same neural network or the first nerves network and second mind
It is different neural networks through network；Detection module 504, for according to Haar algorithm, detection to be located at mobile terminal taking lens
Image in coverage whether include user tongue image；Cue module 505, for working as the mobile terminal taking lens
Coverage in image do not include user tongue image, prompt user adjust mobile terminal taking lens position, directly
Image in the coverage of the mobile terminal taking lens includes the tongue image of user；Image processing module 506 is used
In to the tongue image and tongue sample image progress image cleaning, normalization, dimensionality reduction and/or whitening processing.
Each module of above-mentioned apparatus is corresponding with a kind of lingual diagnosis method of embodiment, and the concrete function of each module can be with reference implementation
A kind of lingual diagnosis method of example, details are not described herein.
The embodiment of the present invention identifies the physical condition of user by locating module and identification module, and wisdom tongue may be implemented
It examines；In addition, using detection module detection customer mobile terminal taking lens coverage in whether the tongue figure containing user
Picture can improve the efficiency of locating module, and the embodiment of the present invention before locating module to the tongue image Primary Location of user
It can be used for mobile terminal, it is user-friendly.
The embodiment of the present application provides a kind of nonvolatile computer storage media, and the computer storage medium is stored with
One of above-mentioned any means embodiment lingual diagnosis method pair can be performed in an at least executable instruction, the computer executable instructions
The operation answered.
Fig. 6 is the structural schematic diagram that the present invention calculates apparatus embodiments, and the specific embodiment of the invention is not to calculating equipment
Specific implementation limit.
As shown in fig. 6, the calculating equipment may include: processor (processor) 602, communication interface
(Communications Interface) 604, memory (memory) 606 and communication bus 608.
Wherein:
Processor 602, communication interface 604 and memory 606 complete mutual communication by communication bus 608.
Communication interface 604, for being communicated with the network element of other equipment such as client or other servers etc..
Processor 602 can specifically execute the correlation in a kind of above-mentioned lingual diagnosis embodiment of the method for executing program 610
Step.
Specifically, program 610 may include program code, which includes computer operation instruction.
Processor 602 may be central processor CPU or specific integrated circuit ASIC (Application
Specific Integrated Circuit), or be arranged to implement the integrated electricity of one or more of the embodiment of the present invention
Road.The one or more processors that equipment includes are calculated, can be same type of processor, such as one or more CPU；It can also
To be different types of processor, such as one or more CPU and one or more ASIC.
Memory 606, for storing program 610.Memory 606 may include high speed RAM memory, it is also possible to further include
Nonvolatile memory (non-volatile memory), for example, at least a magnetic disk storage.
Program 610 specifically can be used for so that processor 602 executes following operation:
Obtain the tongue image of user；After tongue image input positioning neural network is obtained the positioning of tongue target
Tongue image is positioned, the positioning neural network is obtained by first nerves network training, the first nerves network packet
Include SSD neural network；The positioning tongue image is inputted into identification neural network and obtains recognition result, the identification neural network
It is to be obtained by nervus opticus network training, the first nerves network and the nervus opticus network are same nerve net
Network or the first nerves network are different neural networks from the nervus opticus network.
In a kind of optional mode, program 610 can specifically be further used for so that processor 602 executes following behaviour
Make: the method also includes:
Obtain the tongue sample image of preset quantity；The circumscribed frame for marking tongue position in the tongue sample image, obtains
To positioning tongue sample image；By the tongue sample image and positioning tongue sample image input default first mind
Through network, the positioning neural network is obtained.
In a kind of optional mode, program 410 can specifically be further used for so that processor 402 executes following behaviour
Make: obtaining the positioning tongue sample image；Tongue label is added to the positioning tongue sample image, the tongue label is used
In the physical condition for indicating user corresponding with the tongue sample image；By the positioning tongue sample image and with the positioning
The corresponding tongue label of tongue sample image is input to the nervus opticus network and is trained to obtain identification neural network.
In a kind of optional mode, program 610 can specifically be further used for so that processor 602 executes following behaviour
Make: according to Haar algorithm, detection be located at the image in the coverage of mobile terminal taking lens whether include user tongue
Image；If comprising obtaining the tongue image of user.If the image in the coverage of the mobile terminal taking lens does not include
The tongue image of user prompts user to adjust the position of mobile terminal taking lens, until the mobile terminal taking lens
Image in coverage includes the tongue image of user.
In a kind of optional mode, program 610 can specifically be further used for so that processor 602 executes following behaviour
Make: image cleaning, normalization, dimensionality reduction and/or whitening processing are carried out to the tongue image and the tongue sample image.
Algorithm and display are not inherently related to any particular computer, virtual system, or other device provided herein.
Various general-purpose systems can also be used together with teachings based herein.As described above, it constructs required by this kind of system
Structure be obvious.In addition, the present invention is also not directed to any particular programming language.It should be understood that can use various
Programming language realizes summary of the invention described herein, and the description done above to language-specific is to disclose this hair
Bright most preferred embodiment.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects,
Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect
Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, such as right
As claim reflects, inventive aspect is all features less than single embodiment disclosed above.Therefore, it then follows tool
Thus claims of body embodiment are expressly incorporated in the specific embodiment, wherein each claim itself is used as this hair
Bright separate embodiments.
Those skilled in the art will understand that can be carried out adaptively to the module in the equipment in embodiment
Change and they are arranged in one or more devices different from this embodiment.It can be the module or list in embodiment
Member or component are combined into a module or unit or component, and furthermore they can be divided into multiple submodule or subelement or
Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it can use any
Combination is to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed
All process or units of what method or apparatus are combined.Unless expressly stated otherwise, this specification is (including adjoint power
Benefit require, abstract and attached drawing) disclosed in each feature can carry out generation with an alternative feature that provides the same, equivalent, or similar purpose
It replaces.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention
Within the scope of and form different embodiments.For example, in the following claims, embodiment claimed is appointed
Meaning one of can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) come realize some in a kind of lingual diagnosis device according to an embodiment of the present invention or
The some or all functions of person's whole component.The present invention is also implemented as one for executing method as described herein
Point or whole device or device programs (for example, computer program and computer program product).Such this hair of realization
Bright program can store on a computer-readable medium, or may be in the form of one or more signals.It is such
Signal can be downloaded from an internet website to obtain, and is perhaps provided on the carrier signal or is provided in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
Claims (9)
1. a kind of lingual diagnosis method characterized by comprising
Obtain the tongue image of user；
The tongue image is inputted into the positioning tongue image after positioning neural network obtains the positioning of tongue target, the positioning mind
It through network is obtained by first nerves network training, the first nerves network includes SSD neural network；
The positioning tongue image is inputted into identification neural network and obtains recognition result, the identification neural network is by second
What neural metwork training obtained, the first nerves network and the nervus opticus network are same neural network or described
First nerves network is different neural networks from the nervus opticus network.
2. the method according to claim 1, wherein the positioning neural network is by first nerves net
Network training obtains, comprising:
The positioning neural network is obtained to the first nerves network training；
It is described that the positioning neural network is obtained to the first nerves network training, comprising:
Obtain the tongue sample image of preset quantity；
The circumscribed frame for marking tongue position in the tongue sample image obtains positioning tongue sample image；
The tongue sample image and the positioning tongue sample image are inputted the first nerves network to be trained, obtained
The positioning neural network.
3. according to the method described in claim 2, it is characterized in that, the identification neural network is by nervus opticus
What network training obtained, comprising:
The identification neural network is obtained to the nervus opticus network training；
It is described that the identification neural network is obtained to the nervus opticus network training, comprising:
Obtain the positioning tongue sample image；
Tongue label is added to the positioning tongue sample image, the tongue label is used to indicate and the tongue sample image
The physical condition of corresponding user；
The positioning tongue sample image and the tongue label corresponding with the positioning tongue sample image are input to institute
Nervus opticus network is stated to be trained to obtain identification neural network.
4. the method according to claim 1, wherein the method is also before the tongue image for obtaining user
Include:
According to Haar algorithm, detection be located at the image in the coverage of mobile terminal taking lens whether include user tongue
Image；
If comprising obtaining the tongue image of user.
5. according to the method described in claim 4, it is characterized in that, the method also includes:
If the image in the coverage of the mobile terminal taking lens does not include the tongue image of user, user's adjustment is prompted
The position of mobile terminal taking lens, until the image in the coverage of the mobile terminal taking lens includes the tongue of user
Head image.
6. method according to claim 1-4, which is characterized in that the method also includes: to the tongue figure
Picture and the tongue sample image carry out image cleaning, normalization, dimensionality reduction and/or whitening processing.
7. a kind of wisdom lingual diagnosis device based on mobile terminal, which is characterized in that described device includes:
Obtain module: for obtaining the tongue image of user；
Locating module: for the tongue image to be inputted the positioning tongue figure after positioning neural network obtains the positioning of tongue target
Picture, the positioning neural network is obtained by first nerves network training, and the first nerves network includes SSD nerve net
Network；
Identification module: recognition result, the identification nerve are obtained for the positioning tongue image to be inputted identification neural network
Network is obtained by nervus opticus network training, and the first nerves network and the nervus opticus network are same nerve
Network or the first nerves network are different neural networks from the nervus opticus network.
8. a kind of calculating equipment, comprising: processor, memory, communication interface and communication bus, the processor, the storage
Device and the communication interface complete mutual communication by the communication bus；
The memory executes the processor as right is wanted for storing an at least executable instruction, the executable instruction
Ask a kind of corresponding operation of lingual diagnosis method described in any one of 1-6.
9. a kind of computer storage medium, an at least executable instruction, the executable instruction are stored in the storage medium
Processor is set to execute a kind of corresponding operation of lingual diagnosis method such as of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811634491.XA CN109712707A (en) | 2018-12-29 | 2018-12-29 | A kind of lingual diagnosis method, apparatus calculates equipment and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811634491.XA CN109712707A (en) | 2018-12-29 | 2018-12-29 | A kind of lingual diagnosis method, apparatus calculates equipment and computer storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109712707A true CN109712707A (en) | 2019-05-03 |
Family
ID=66258234
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811634491.XA Pending CN109712707A (en) | 2018-12-29 | 2018-12-29 | A kind of lingual diagnosis method, apparatus calculates equipment and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109712707A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110164550A (en) * | 2019-05-22 | 2019-08-23 | 杭州电子科技大学 | A kind of congenital heart disease aided diagnosis method based on multi-angle of view conspiracy relation |
| CN110363072A (en) * | 2019-05-31 | 2019-10-22 | 正和智能网络科技(广州)有限公司 | Tongue image recognition method, apparatus, computer equipment and computer readable storage medium |
| CN113362334A (en) * | 2020-03-04 | 2021-09-07 | 北京悦熙兴中科技有限公司 | Tongue picture processing method and device |
| CN113837987A (en) * | 2020-12-31 | 2021-12-24 | 京东科技控股股份有限公司 | Tongue image acquisition method and device and computer equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103905733A (en) * | 2014-04-02 | 2014-07-02 | 哈尔滨工业大学深圳研究生院 | Method and system for conducting real-time tracking on faces by monocular camera |
| CN105357442A (en) * | 2015-11-27 | 2016-02-24 | 小米科技有限责任公司 | Shooting angle adjustment method and device for camera |
| CN106295139A (en) * | 2016-07-29 | 2017-01-04 | 姹ゅ钩 | A kind of tongue body autodiagnosis health cloud service system based on degree of depth convolutional neural networks |
| CN108986912A (en) * | 2018-07-12 | 2018-12-11 | 北京三医智慧科技有限公司 | Chinese medicine stomach trouble tongue based on deep learning is as information intelligent processing method |
-
2018
- 2018-12-29 CN CN201811634491.XA patent/CN109712707A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103905733A (en) * | 2014-04-02 | 2014-07-02 | 哈尔滨工业大学深圳研究生院 | Method and system for conducting real-time tracking on faces by monocular camera |
| CN105357442A (en) * | 2015-11-27 | 2016-02-24 | 小米科技有限责任公司 | Shooting angle adjustment method and device for camera |
| CN106295139A (en) * | 2016-07-29 | 2017-01-04 | 姹ゅ钩 | A kind of tongue body autodiagnosis health cloud service system based on degree of depth convolutional neural networks |
| CN108986912A (en) * | 2018-07-12 | 2018-12-11 | 北京三医智慧科技有限公司 | Chinese medicine stomach trouble tongue based on deep learning is as information intelligent processing method |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110164550A (en) * | 2019-05-22 | 2019-08-23 | 杭州电子科技大学 | A kind of congenital heart disease aided diagnosis method based on multi-angle of view conspiracy relation |
| CN110164550B (en) * | 2019-05-22 | 2021-07-09 | 杭州电子科技大学 | Congenital heart disease auxiliary diagnosis method based on multi-view cooperative relationship |
| CN110363072A (en) * | 2019-05-31 | 2019-10-22 | 正和智能网络科技(广州)有限公司 | Tongue image recognition method, apparatus, computer equipment and computer readable storage medium |
| CN110363072B (en) * | 2019-05-31 | 2023-06-09 | 正和智能网络科技(广州)有限公司 | Tongue picture identification method, tongue picture identification device, computer equipment and computer readable storage medium |
| CN113362334A (en) * | 2020-03-04 | 2021-09-07 | 北京悦熙兴中科技有限公司 | Tongue picture processing method and device |
| CN113362334B (en) * | 2020-03-04 | 2024-05-24 | 北京悦熙兴中科技有限公司 | Tongue photo processing method and device |
| CN113837987A (en) * | 2020-12-31 | 2021-12-24 | 京东科技控股股份有限公司 | Tongue image acquisition method and device and computer equipment |
| CN113837987B (en) * | 2020-12-31 | 2023-11-03 | 京东科技控股股份有限公司 | Tongue image acquisition method and device and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109712707A (en) | A kind of lingual diagnosis method, apparatus calculates equipment and computer storage medium | |
| CN110232383B (en) | Focus image recognition method and focus image recognition system based on deep learning model | |
| CN109964235A (en) | For carrying out the prediction model of vision sorter to insect | |
| CN109829446A (en) | Eye fundus image recognition methods, device, electronic equipment and storage medium | |
| CN108875934A (en) | A kind of training method of neural network, device, system and storage medium | |
| CN109376663A (en) | A kind of human posture recognition method and relevant apparatus | |
| CN110472737A (en) | Training method, device and the magic magiscan of neural network model | |
| CN106682127A (en) | Image searching system and method | |
| CN115345938B (en) | Global-to-local-based head shadow mark point positioning method, equipment and medium | |
| CN113516639B (en) | Training method and device for oral cavity abnormality detection model based on panoramic X-ray film | |
| CN111681247A (en) | Lung lobe and lung segment segmentation model training method and device | |
| CN112614573A (en) | Deep learning model training method and device based on pathological image labeling tool | |
| CN110742690A (en) | Method for configuring endoscope and terminal equipment | |
| Raja et al. | Convolutional Neural Networks based Classification and Detection of Plant Disease | |
| US10970863B2 (en) | System and method of analyzing features of the human face and breasts using one or more overlay grids | |
| Sapkota et al. | Comparing YOLO11 and YOLOv8 for instance segmentation of occluded and non-occluded immature green fruits in complex orchard environment | |
| CN110110750A (en) | A kind of classification method and device of original image | |
| Minowa et al. | Convolutional neural network applied to tree species identification based on leaf images | |
| Pho et al. | Segmentation-driven hierarchical retinanet for detecting protozoa in micrograph | |
| CN112668668B (en) | Postoperative medical image evaluation method and device, computer equipment and storage medium | |
| CN114495097A (en) | Multi-model-based urine cell identification method and system | |
| CN110110749A (en) | Image processing method and device in a kind of training set | |
| Aleynikov et al. | Automated diseases detection of plant diseases in space greenhouses | |
| Dhar et al. | A Deep Learning Approach to Teeth Segmentation and Orientation from Panoramic X-rays | |
| Parakh et al. | Detection of Bell Pepper Crop Diseases Using Convolution Neural Network |
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: 20190503 |
|
| RJ01 | Rejection of invention patent application after publication |