CN110334763A - Model data file generation, image-recognizing method, device, equipment and medium - Google Patents
Model data file generation, image-recognizing method, device, equipment and medium Download PDFInfo
- Publication number
- CN110334763A CN110334763A CN201910599358.3A CN201910599358A CN110334763A CN 110334763 A CN110334763 A CN 110334763A CN 201910599358 A CN201910599358 A CN 201910599358A CN 110334763 A CN110334763 A CN 110334763A
- Authority
- CN
- China
- Prior art keywords
- model data
- data file
- image
- model
- client
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
- G06V10/464—Salient features, e.g. scale invariant feature transforms [SIFT] using a plurality of salient features, e.g. bag-of-words [BoW] representations
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present disclosure discloses a kind of model data file generation, image-recognizing method, device, equipment and media.It is white space in addition to target object in individual described training image this method comprises: being trained according to individual training image to initial machine learning model；It is sent to client with the matched model data file of the target object by what training obtained, so that client is based on the model data file and identifies the target object.The embodiment of the present disclosure can accelerate model formation speed, shorten the model training period.
Description
Technical field
The embodiment of the present disclosure is related to data processing technique more particularly to a kind of generation of model data file, image recognition side
Method, device, equipment and medium.
Background technique
With the development of machine learning techniques, machine learning model has become common and accurate image-recognizing method.
For example, can be input to the image of acquisition in machine learning model trained in advance in field of image recognition,
Obtain the image for being labeled with target object.If thinking, model can recognize that target object, need to collect comprising the target object
Multiple images, and the multiple images not comprising the target object, respectively as training sample, even, it is also necessary to by artificial
It is labeled in each training sample.To be input in model to the training sample marked and carry out model training, will trained
The model of completion comes into operation as image recognition model.
In the method that above-mentioned model generates, training process is complicated, and the training time is very long.
Summary of the invention
The embodiment of the present disclosure provides a kind of model data file generation, image-recognizing method, device, equipment and medium, can
To accelerate model formation speed, shorten the model training period.
In a first aspect, the embodiment of the present disclosure provides a kind of model data file generation method, this method comprises:
Initial machine learning model is trained according to individual training image, target is removed in individual described training image
It is white space except object；
It is sent to client with the matched model data file of the target object by what training obtained, so that client's end group
The target object is identified in the model data file.
Second aspect, the embodiment of the present disclosure additionally provides a kind of image-recognizing method, using in the client, comprising:
When receiving the image recognition instruction for image to be detected, at least one model data file is loaded into interior
In depositing；Wherein, the model data file is by being trained shape to initial machine learning model according to individual training image
At；
Image recognition is carried out to described image to be detected respectively based on each model data file, is obtained described to be detected
The image recognition result of image.
The third aspect, the embodiment of the present disclosure additionally provide a kind of model data file generating means, which includes:
Single sample training module, it is described for being trained according to individual training image to initial machine learning model
It is white space in addition to target object in individual training image；
Model data file generation module, for that will train and the target object matched model data file
It is sent to client, so that client is based on the model data file and identifies the target object.
Fourth aspect, the embodiment of the present disclosure additionally provide a kind of pattern recognition device, which includes:
Model data file loading module, for receive for image to be detected image recognition instruction when, it is near
A few model data file is loaded into memory；Wherein, the model data file by according to individual training image to first
The machine learning model of beginning is trained to be formed；
Picture recognition module, for carrying out image knowledge to described image to be detected respectively based on each model data file
Not, the image recognition result of described image to be detected is obtained.
5th aspect, the embodiment of the present disclosure additionally provide a kind of electronic equipment, which includes:
One or more processors；
Memory, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes model data file generation method as described in the embodiment of the present disclosure is any or as described in the embodiment of the present disclosure is any
Image-recognizing method.
6th aspect, the embodiment of the present disclosure additionally provide a kind of computer readable storage medium, are stored thereon with computer
Program realizes model data file generation method as described in the embodiment of the present disclosure is any or such as when the program is executed by processor
Any image-recognizing method of the embodiment of the present disclosure.
The embodiment of the present disclosure by obtaining matched model data file to individual training image training machine learning model,
The quantity for reducing the training sample of machine learning model improves trained speed to reduce the data volume that training process is related to,
Shorten the model training period, so that model data file formation efficiency is improved, meanwhile, it is operated only for single image, letter
Change training process, the complexity that model generates is reduced, and model data file is issued to client, so that client has
It identifies the function of target object, improves client models and update efficiency, the generating process for solving model in the prior art is very long
And complicated problem, the formation efficiency of model data file is improved, the model training period is shortened, accelerates model formation speed,
The model modification efficiency of client is improved simultaneously.
Detailed description of the invention
Fig. 1 a is the flow chart of one of embodiment of the present disclosure one model data file generation method；
Fig. 1 b is the schematic diagram of individual training image of one of the embodiment of the present disclosure one；
Fig. 2 is the flow chart of one of the embodiment of the present disclosure two image-recognizing method；
Fig. 3 is the flow chart of one of the embodiment of the present disclosure three image-recognizing method；
Fig. 4 is the structural schematic diagram of one of embodiment of the present disclosure four model data file generating means；
Fig. 5 is the structural schematic diagram of one of the embodiment of the present disclosure five pattern recognition device；
Fig. 6 is the structural schematic diagram of one of the embodiment of the present disclosure six electronic equipment.
Specific embodiment
Embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although showing the certain of the disclosure in attached drawing
Embodiment, it should be understood that, the disclosure can be realized by various forms, and should not be construed as being limited to this
In the embodiment that illustrates, providing these embodiments on the contrary is in order to more thorough and be fully understood by the disclosure.It should be understood that
It is that being given for example only property of the accompanying drawings and embodiments effect of the disclosure is not intended to limit the protection scope of the disclosure.
It should be appreciated that each step recorded in disclosed method embodiment can execute in a different order,
And/or parallel execution.In addition, method implementation may include additional step and/or omit the step of execution is shown.This public affairs
The range opened is not limited in this respect.
Terms used herein " comprising " and its deformation are that opening includes, i.e., " including but not limited to ".Term "based"
It is " being based at least partially on ".Term " one embodiment " expression " at least one embodiment "；Term " another embodiment " indicates
" at least one other embodiment "；Term " some embodiments " expression " at least some embodiments ".The correlation of other terms is fixed
Justice provides in will be described below.
It is noted that the concepts such as " first " that refers in the disclosure, " second " are only used for different devices, module or list
Member distinguishes, and is not intended to limit the sequence or relation of interdependence of function performed by these devices, module or unit.
It is noted that referred in the disclosure "one", the modification of " multiple " be schematically and not restrictive this field
It will be appreciated by the skilled person that being otherwise construed as " one or more " unless clearly indicate otherwise in context.
The being merely to illustrate property of title of the message or information that are interacted between multiple devices in disclosure embodiment
Purpose, and be not used to limit the range of these message or information.
Embodiment one
Fig. 1 a is the flow chart of one of the embodiment of the present disclosure one model data file generation method, and the present embodiment can fit
The case where for generating model data file and real time down, this method can be executed by model data file generating means,
The device can realize that the device can be configured in electronic equipment by the way of software and/or hardware, for example, server or
Terminal device, typical terminal device include desktop computer or laptop etc..As shown in Figure 1a, this method specifically includes
Following steps:
S110 is trained initial machine learning model according to individual training image, in individual described training image
It is white space in addition to target object.
Individual training image is used for training machine learning model.Individual training image only includes target object, for only mentioning
The characteristic point for getting target object reduces the interference of the characteristic point of other non-targeted objects, thus, guarantee based on individual training figure
As the recognition accuracy for the machine learning model that training is formed.
Target object can refer to object to be identified.Optionally, the target object includes plane pattern.It is exemplary
, plane pattern includes poster figure or trademark image.It illustratively, as shown in Figure 1 b, only include target object in training image 131
132, other regions are white space, wherein target object 132 is the picture comprising lightning pattern.
Machine learning model, which can refer to, carries out the model that image recognition technology realizes identification based on characteristics of image, exemplary
, machine learning model can be bag of words (Bag of word).
S120 is sent to client with the matched model data file of the target object for what training obtained, so that objective
Family end group identifies the target object in the model data file.
Model data file is for storage and the associated characteristic of target object, for example, model data file includes single
Open the code book of training image and the characteristic point etc. of individual training image.
Optionally, the model data file stores the characteristic of individual training image, in the characteristic
Including characteristic point number be more than setting quantity threshold.
It is understood that if the extractible characteristic point of target object is excessively few namely feature of target object is not shown
It writes, is difficult to distinguish the target object and other objects, it is low so as to cause the recognition accuracy of model.
Setting quantity threshold is used to limit the number of extractible characteristic point in target object, illustratively, sets number
Threshold value is 200.
By configured number threshold value, the quantity of extractible characteristic point in target object is limited, guaranteeing can in target object
To extract the characteristic point for being largely different from other objects, guarantee the mould based on a large amount of characteristics for being stored with the target object
Target object can be recognized accurately in type data file, to improve the recognition accuracy of target object.
In a specific example, machine learning model training process: extracting the characteristic point of the target image,
In, characteristic point number is more than setting quantity threshold；Whole characteristic points are clustered, form multiple classes, and determine each class
Cluster centre；Determine each characteristic point of the target image at a distance from the cluster centre of each class, and based on it is each it is described away from
It is encoded to the target image, obtains the code book of the target image.
Specifically, machine learning model includes bag of words.Feature is extracted from individual training image, ORB can be used
(Oriented FAST and Rotated BRIEF) algorithm, the algorithm are simplified to quick (FAST) feature point detecting method
(BREIF) Feature Descriptor combines.Feature is really the key message for referring to represent individual training image.It wherein, will be from
The whole characteristic points extracted in each training image are clustered, and determine the center of each class, wherein whole characteristic points are formed
Bag of words.It determines that code book can specifically refer to the histogram for seeking the training image, i.e., falls in the characteristic point of the training image often
Quantity in a class, to obtain the code book of feature namely training image of the training image under bag of words.
Optionally, described to be sent to client, comprising: by model data coffret predetermined, by the mould
Type data file is sent to client.
Model data coffret is for carrying out model data file transmission.Before main program publication, need to define mould
Type data transmission interface, model data coffret can not be changed after main program publication, until the hair version period next time,
Main program updates model data coffret when sending out version again.
By pre-defining model data coffret, the accurate and real-time transmission of implementation model data file is improved
The update efficiency of model data file, to increase the identifiable object range of client models.
Optionally, model data file and real time down are generated by server.Specifically, provide for user can for server
Depending on changing the page, the control for adding individual training image is shown in the visual page.Specifically, when user adds individual
After training image, server Auto-generation Model data file, and real time down is to each client.In addition, the visualization page
Face further includes at least one of following: the preview area of the thumbnail of individual training image, the address of individual training image, mould
Type training control and control etc. for issuing the model data file of generation.The visual page can also include it
His content, in this regard, the embodiment of the present disclosure is not specifically limited.
The embodiment of the present disclosure by obtaining matched model data file to individual training image training machine learning model,
The quantity for reducing the training sample of machine learning model improves trained speed to reduce the data volume that training process is related to,
Shorten the model training period, so that model data file formation efficiency is improved, meanwhile, it is operated only for single image, letter
Change training process, the complexity that model generates is reduced, and model data file is issued to client, so that client has
It identifies the function of target object, improves client models and update efficiency, the generating process for solving model in the prior art is very long
And complicated problem, the formation efficiency of model data file is improved, the model training period is shortened, accelerates model formation speed,
The model modification efficiency of client is improved simultaneously.
Embodiment two
Fig. 2 is the flow chart of one of the embodiment of the present disclosure two image-recognizing method, and the present embodiment is applicable to treat
Detection image carries out the case where image recognition, and this method can be executed by pattern recognition device, which can use software
And/or the mode of hardware is realized, which can be configured in electronic equipment, such as mobile terminal, typical mobile terminal packet
Include mobile phone, car-mounted terminal or laptop etc..This method specifically comprises the following steps:
S210 adds at least one model data file when receiving the image recognition instruction for image to be detected
It is downloaded in memory；Wherein, the model data file is by carrying out initial machine learning model according to individual training image
Training is formed.
Image to be detected can be pre-stored image, the video frame being also possible in the video of real-time recording.Image
Identification instruction is for stress model data file and calls image identification model function program.Wherein, image recognition instruction can
To refer to instruction that user is inputted by trigger action.
In a specific example, user can sweep control or other image recognition controls by touching to sweep, and call
Camera carries out the captured in real-time of video, acquires image to be detected, while being measured in real time to image to be detected.
Specifically, image recognition model function program can be a main program；Or it can be one and be with main program
The program of running environment operation.
Model data file can refer to the description of above-described embodiment.
After model data file is loaded into memory, the available load of image recognition model function program is in memory
Model data file in data, and realize image identification function.
S220 carries out image recognition to described image to be detected respectively based on each model data file, obtains described
The image recognition result of image to be detected.
In fact, different model data files respectively correspond the characteristic for being stored with different known objects, image recognition
Model function program can be matched the characteristic in model data file one by one with image to be detected respectively, will be matched
Spend high conduct image recognition result.
Specifically, as in the previous example, machine learning model is bag of words, image recognition model function program extracts mapping to be checked
The characteristic point of picture, and the existing class based on bag of words carry out clustering, obtain the image to be detected under bag of words
Target feature point.The characteristic point for obtaining each known object calculates separately the spy of each target feature point Yu each known object
The distance for levying point, determines that image to be detected, will be apart from the smallest known object as to be detected at a distance from each known object
The image recognition result of image.
Optionally, before receiving for the image recognition instruction of target image, further includes: transmitted by model data
Interface receives the model data file that server issues.
Model data coffret is for carrying out model data file transmission.Specifically, model data coffret is used for
Model data transmission service is provided.By pre-defining model data coffret, implementation model data file accurate and
Real-time reception improves the update efficiency of model data file, and carries out image recognition according to updated model data file,
Increase the identifiable object range of client models, so as to identify the unrecognized object of history, improves object and know
Other accuracy rate.
The model data file that the embodiment of the present disclosure is obtained by obtaining individual training image training machine learning model, subtracts
The training sample of few machine learning model reduces amount of training data and simplifies training process, accelerates obtaining for model data file
Speed is taken, and improves model data file and updates efficiency, and image recognition is carried out according to updated model data file,
Increase the identifiable object range of client models, so as to identify the unrecognized object of history, improves object and know
Other accuracy rate.
Embodiment three
Fig. 3 is the flow chart of one of the embodiment of the present disclosure three image-recognizing method, present embodiments provides server
With the interaction scenario of client, it is suitable for above-described embodiment.
This method specifically comprises the following steps:
S310, server are trained initial machine learning model according to individual training image, individual described training
It is white space in addition to target object in image；
Individual training image, machine learning model, model data coffret, model data text in the embodiment of the present disclosure
Part and image to be detected etc. can refer to the description of above-described embodiment.
S320, the server by model data coffret predetermined will training it is obtaining with the object
The matched model data file of body is sent to client.
S330, the client receive the model data file by the model data coffret.
S340, when receiving the image recognition instruction for image to be detected, the client is by least one model
Data file is loaded into memory.
S350, the client are based on each model data file and carry out image knowledge to described image to be detected respectively
Not, the image recognition result of described image to be detected is obtained.
The embodiment of the present disclosure by obtaining matched model data file to individual training image training machine learning model,
The training sample of machine learning model is reduced, amount of training data is reduced and simplifies training process, accelerates model data file
Formation speed, meanwhile, by model data file real time down into client, improves client models and update efficiency, Er Qieke
Family end can carry out image recognition according to updated model data file, increase the identifiable object model of client models
It encloses, so as to identify the unrecognized object of history, improves the accuracy rate of object identification.
Example IV
Fig. 4 is a kind of structural schematic diagram for model data file generating means that the embodiment of the present disclosure four provides, this implementation
Example is applicable to the case where generating model data file and real time down.The device can be by the way of software and/or hardware
It realizes, which can be configured in electronic equipment, such as server or terminal device, and typical terminal device includes desk-top meter
Calculation machine or laptop etc..As shown in figure 4, the apparatus may include: single sample training module 410 and model data file are raw
At module 420.
Single sample training module 410, for being trained according to individual training image to initial machine learning model, institute
Stating in individual training image is white space in addition to target object；
Model data file generation module 420, for that will train and the target object matched model data
File is sent to client, so that client is based on the model data file and identifies the target object.
The embodiment of the present disclosure by obtaining matched model data file to individual training image training machine learning model,
The quantity for reducing the training sample of machine learning model improves trained speed to reduce the data volume that training process is related to,
Shorten the model training period, so that model data file formation efficiency is improved, meanwhile, it is operated only for single image, letter
Change training process, the complexity that model generates is reduced, and model data file is issued to client, so that client has
It identifies the function of target object, improves client models and update efficiency, the generating process for solving model in the prior art is very long
And complicated problem, the formation efficiency of model data file is improved, the model training period is shortened, accelerates model formation speed,
The model modification efficiency of client is improved simultaneously.
Further, the model data file stores the characteristic of individual training image, the characteristic
In include characteristic point number be more than setting quantity threshold.
Further, the model data file generation module 420, comprising: model data coffret transmission unit is used
In by model data coffret predetermined, the model data file is sent to client.
Further, the target object includes plane pattern.
The model data file generating means that the embodiment of the present disclosure provides, the model data file provided with previous embodiment
Generation method belongs to same inventive concept, and the technical detail of detailed description not can be found in aforementioned implementation in the embodiments of the present disclosure
Example, and the model data file generating means and the model data file life of previous embodiment offer that the embodiment of the present disclosure provides
At method beneficial effect having the same.
Embodiment five
Fig. 5 is a kind of structural schematic diagram for pattern recognition device that the embodiment of the present disclosure five provides, and the present embodiment is applicable
In to image to be detected carry out image recognition the case where.The device can realize by the way of software and/or hardware, the device
It can be configured in electronic equipment, such as mobile terminal, typical mobile terminal includes mobile phone, car-mounted terminal or laptop
Deng.As shown in figure 5, the apparatus may include: model data file loading module 510 and picture recognition module 520.
Model data file loading module 510, for inciting somebody to action when receiving the image recognition instruction for image to be detected
At least one model data file is loaded into memory；Wherein, the model data file passes through according to individual training image pair
Initial machine learning model is trained to be formed；
Picture recognition module 520, for carrying out figure to described image to be detected respectively based on each model data file
As identification, the image recognition result of described image to be detected is obtained.
The model data file that the embodiment of the present disclosure is obtained by obtaining individual training image training machine learning model, subtracts
The training sample of few machine learning model reduces amount of training data and simplifies training process, accelerates obtaining for model data file
Speed is taken, and improves model data file and updates efficiency, and image recognition is carried out according to updated model data file,
Increase the identifiable object range of client models, so as to identify the unrecognized object of history, improves object and know
Other accuracy rate.
Further, described image identification device further include: model data coffret transmission module, for receiving
Before the image recognition instruction of target image, by model data coffret, the model data that server issues is received
File.
The pattern recognition device that the embodiment of the present disclosure provides belongs to same with the image-recognizing method that previous embodiment provides
Inventive concept, the technical detail of detailed description not can be found in previous embodiment in the embodiments of the present disclosure, and the disclosure is implemented
The image-recognizing method beneficial effect having the same that the pattern recognition device and previous embodiment that example provides provide.
Embodiment six
Below with reference to Fig. 6, it illustrates the electronic equipment for being suitable for being used to realize the embodiment of the present disclosure (such as server or ends
End equipment) 600 structural schematic diagram.Terminal device in the embodiment of the present disclosure can include but is not limited to such as mobile phone,
Laptop, digit broadcasting receiver, PDA (personal digital assistant), PAD (tablet computer), PMP (portable multimedia broadcasting
Put device), the mobile terminal of car-mounted terminal (such as vehicle mounted guidance terminal) etc. and such as number TV, desktop computer etc.
Fixed terminal.Electronic equipment shown in Fig. 6 is only an example, should not function and use scope band to the embodiment of the present disclosure
Carry out any restrictions.
As shown in fig. 6, electronic equipment 600 may include processing unit (such as central processing unit, graphics processor etc.)
601, random access can be loaded into according to the program being stored in read-only memory (ROM) 602 or from storage device 608
Program in memory (RAM) 603 and execute various movements appropriate and processing.In RAM 603, it is also stored with electronic equipment
Various programs and data needed for 600 operations.Processing unit 601, ROM 602 and RAM 603 pass through the phase each other of bus 604
Even.Input/output (I/O) interface 605 is also connected to bus 604.
In general, following device can connect to I/O interface 605: including such as touch screen, touch tablet, keyboard, mouse, taking the photograph
As the input unit 606 of head, microphone, accelerometer, gyroscope etc.；Including such as liquid crystal display (LCD), loudspeaker, vibration
The output device 607 of dynamic device etc.；Storage device 608 including such as tape, hard disk etc.；And communication device 609.Communication device
609, which can permit electronic equipment 600, is wirelessly or non-wirelessly communicated with other equipment to exchange data.Although Fig. 6 shows tool
There is the electronic equipment 600 of various devices, it should be understood that being not required for implementing or having all devices shown.It can be with
Alternatively implement or have more or fewer devices.
Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description
Software program.For example, embodiment of the disclosure includes a kind of computer program product comprising being carried on non-transient computer can
The computer program on medium is read, which includes the program code for method shown in execution flow chart.At this
In the embodiment of sample, which can be downloaded and installed from network by communication device 609, or be filled from storage
It sets 608 to be mounted, or is mounted from ROM 602.When the computer program is executed by processing unit 601, the disclosure is executed
The above-mentioned function of being limited in the method for embodiment.
Embodiment seven
The above-mentioned computer-readable medium of the disclosure can be computer-readable signal media or computer-readable storage
Medium either the two any combination.Computer readable storage medium for example may be-but not limited to-electricity,
Magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or any above combination.Computer-readable storage
The more specific example of medium can include but is not limited to: have electrical connection, the portable computer magnetic of one or more conducting wires
Disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable programmable read only memory (EPROM or sudden strain of a muscle
Deposit), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device or above-mentioned appoint
The suitable combination of meaning.In the disclosure, computer readable storage medium can be any tangible medium for including or store program,
The program can be commanded execution system, device or device use or in connection.And in the disclosure, computer
Readable signal medium may include in a base band or as the data-signal that carrier wave a part is propagated, wherein carrying computer
Readable program code.The data-signal of this propagation can take various forms, including but not limited to electromagnetic signal, optical signal
Or above-mentioned any appropriate combination.Computer-readable signal media can also be any other than computer readable storage medium
Computer-readable medium, the computer-readable signal media can be sent, propagated or transmitted for by instruction execution system, dress
It sets or device uses or program in connection.The program code for including on computer-readable medium can be with any
Medium transmission appropriate, including but not limited to: electric wire, optical cable, RF (radio frequency) etc. or above-mentioned any appropriate combination.
In some embodiments, client, server can use such as HTTP (HyperText Transfer
Protocol, hypertext transfer protocol) etc the network protocols of any currently known or following research and development communicated, and can
To be interconnected with the digital data communications (for example, communication network) of arbitrary form or medium.The example of communication network includes local area network
(" LAN "), wide area network (" WAN "), Internet (for example, internet) and ad-hoc network are (for example, the end-to-end net of ad hoc
Network) and any currently known or following research and development network.
Above-mentioned computer-readable medium can be included in above-mentioned electronic equipment；It is also possible to individualism, and not
It is fitted into the electronic equipment.
Above-mentioned computer-readable medium carries one or more program, when said one or multiple programs are by the electricity
When sub- equipment executes, so that the electronic equipment: initial machine learning model is trained according to individual training image, it is described
It is white space in addition to target object in individual training image；By training obtain with the matched pattern number of the target object
It is sent to client according to file, so that client is based on the model data file and identifies the target object.
Either when said one or multiple programs are executed by the electronic equipment, so that the electronic equipment: receiving
For image to be detected image recognition instruction when, at least one model data file is loaded into memory；Wherein, the mould
Type data file to be formed by being trained according to individual training image to initial machine learning model；Based on each model
Data file carries out image recognition to described image to be detected respectively, obtains the image recognition result of described image to be detected.
The calculating of the operation for executing the disclosure can be write with one or more programming languages or combinations thereof
Machine program code, above procedure design language include but is not limited to object oriented program language-such as Java,
Smalltalk, C++ further include conventional procedural programming language-such as " C " language or similar program design language
Speech.Program code can be executed fully on the user computer, partly be executed on the user computer, as an independence
Software package execute, part on the user computer part execute on the remote computer or completely in remote computer or
It is executed on server.In situations involving remote computers, remote computer can pass through the network of any kind --- packet
It includes local area network (LAN) or wide area network (WAN)-is connected to subscriber computer, or, it may be connected to outer computer (such as benefit
It is connected with ISP by internet).
Flow chart and block diagram in attached drawing are illustrated according to the system of the various embodiments of the disclosure, method and computer journey
The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation
A part of one module, program segment or code of table, a part of the module, program segment or code include one or more use
The executable instruction of the logic function as defined in realizing.It should also be noted that in some implementations as replacements, being marked in box
The function of note can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are actually
It can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it to infuse
Meaning, the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart can be with holding
The dedicated hardware based system of functions or operations as defined in row is realized, or can use specialized hardware and computer instruction
Combination realize.
Being described in module involved in the embodiment of the present disclosure can be realized by way of software, can also be by hard
The mode of part is realized.Wherein, the title of module does not constitute the restriction to the module itself under certain conditions, for example, single
Sample training module be also described as " initial machine learning model is trained according to individual training image, it is described
Module in individual training image in addition to target object for white space ".
Function described herein can be executed at least partly by one or more hardware logic components.Example
Such as, without limitation, the hardware logic component for the exemplary type that can be used include: field programmable gate array (FPGA), specially
With integrated circuit (ASIC), Application Specific Standard Product (ASSP), system on chip (SOC), complex programmable logic equipment (CPLD) etc.
Deng.
In the context of the disclosure, machine readable media can be tangible medium, may include or is stored for
The program that instruction execution system, device or equipment are used or is used in combination with instruction execution system, device or equipment.Machine can
Reading medium can be machine-readable signal medium or machine-readable storage medium.Machine readable media can include but is not limited to electricity
Son, magnetic, optical, electromagnetism, infrared or semiconductor system, device or equipment or above content any conjunction
Suitable combination.The more specific example of machine readable storage medium will include the electrical connection of line based on one or more, portable meter
Calculation machine disk, hard disk, random access memory (RAM), read-only memory (ROM), Erasable Programmable Read Only Memory EPROM (EPROM
Or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage facilities or
Any appropriate combination of above content.
According to one or more other embodiments of the present disclosure, present disclose provides a kind of model data file generation method, packets
It includes:
Initial machine learning model is trained according to individual training image, target is removed in individual described training image
It is white space except object；
It is sent to client with the matched model data file of the target object by what training obtained, so that client's end group
The target object is identified in the model data file.
It is described in the model data file generation method provided according to one or more other embodiments of the present disclosure, the disclosure
Model data file stores the characteristic of individual training image, and the number for the characteristic point for including in the characteristic is super
Cross setting quantity threshold.
It is described in the model data file generation method provided according to one or more other embodiments of the present disclosure, the disclosure
It is sent to client, comprising: by model data coffret predetermined, the model data file is sent to client
End.
It is described in the model data file generation method provided according to one or more other embodiments of the present disclosure, the disclosure
Target object includes plane pattern.
According to one or more other embodiments of the present disclosure, present disclose provides a kind of image-recognizing methods, comprising:
When receiving the image recognition instruction for image to be detected, at least one model data file is loaded into interior
In depositing；Wherein, the model data file is by being trained shape to initial machine learning model according to individual training image
At；
Image recognition is carried out to described image to be detected respectively based on each model data file, is obtained described to be detected
The image recognition result of image.
In the image-recognizing method provided according to one or more other embodiments of the present disclosure, the disclosure, it is directed to receiving
Before the image recognition instruction of target image, further includes: by model data coffret, receive the pattern number that server issues
According to file.
According to one or more other embodiments of the present disclosure, present disclose provides a kind of model data file generating means, packets
It includes:
Single sample training module, it is described for being trained according to individual training image to initial machine learning model
It is white space in addition to target object in individual training image；
Model data file generation module, for that will train and the target object matched model data file
It is sent to client, so that client is based on the model data file and identifies the target object.
It is described in the model data file generating means provided according to one or more other embodiments of the present disclosure, the disclosure
Model data file stores the characteristic of individual training image, and the number for the characteristic point for including in the characteristic is super
Cross setting quantity threshold.
It is described in the model data file generating means provided according to one or more other embodiments of the present disclosure, the disclosure
Model data file generation module, comprising: model data coffret transmission unit, for passing through model data predetermined
The model data file is sent to client by coffret.
According to the model data file generating means that one or more other embodiments of the present disclosure, the disclosure provide, the mesh
Marking object includes plane pattern.
According to one or more other embodiments of the present disclosure, present disclose provides a kind of pattern recognition devices, are configured at client
In end, comprising:
Model data file loading module, for receive for image to be detected image recognition instruction when, it is near
A few model data file is loaded into memory；Wherein, the model data file by according to individual training image to first
The machine learning model of beginning is trained to be formed；
Picture recognition module, for carrying out image knowledge to described image to be detected respectively based on each model data file
Not, the image recognition result of described image to be detected is obtained.
The pattern recognition device provided according to one or more other embodiments of the present disclosure, the disclosure, further includes: model data
Coffret transmission module, for being transmitted by model data before receiving for the image recognition instruction of target image
Interface receives the model data file that server issues.
According to one or more other embodiments of the present disclosure, present disclose provides a kind of electronic equipment, comprising:
One or more processors；
Memory, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realize the disclosure provide it is any as described in model data file generation method or such as the embodiment of the present disclosure it is any as described in
Image-recognizing method.
According to one or more other embodiments of the present disclosure, present disclose provides a kind of computer readable storage mediums, thereon
It is stored with computer program, realizes when which is executed by processor and is given birth to the disclosure provides any model data file
At method or the image-recognizing method as described in the embodiment of the present disclosure is any.
Above description is only the preferred embodiment of the disclosure and the explanation to institute's application technology principle.Those skilled in the art
Member is it should be appreciated that the open scope involved in the disclosure, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic
Scheme, while should also cover in the case where not departing from design disclosed above, it is carried out by above-mentioned technical characteristic or its equivalent feature
Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed in the disclosure
Can technical characteristic replaced mutually and the technical solution that is formed.
Although this is not construed as requiring these operations with institute in addition, depicting each operation using certain order
The certain order that shows executes in sequential order to execute.Under certain environment, multitask and parallel processing may be advantageous
's.Similarly, although containing several specific implementation details in being discussed above, these are not construed as to this public affairs
The limitation for the range opened.Certain features described in the context of individual embodiment can also be realized in combination single real
It applies in example.On the contrary, the various features described in the context of single embodiment can also be individually or with any suitable
The mode of sub-portfolio is realized in various embodiments.
Although having used specific to this theme of the language description of structure feature and/or method logical action, answer
When understanding that theme defined in the appended claims is not necessarily limited to special characteristic described above or movement.On on the contrary,
Special characteristic described in face and movement are only to realize the exemplary forms of claims.
Claims (10)
1. a kind of model data file generation method characterized by comprising
Initial machine learning model is trained according to individual training image, target object is removed in individual described training image
Except be white space；
It is sent to client with the matched model data file of the target object by what training obtained, so that client is based on institute
It states model data file and identifies the target object.
2. the method according to claim 1, wherein the model data file stores individual described training image
Characteristic, the number for the characteristic point for including in the characteristic is more than setting quantity threshold.
3. the method according to claim 1, wherein described be sent to client, comprising:
By model data coffret predetermined, the model data file is sent to client.
4. the method according to claim 1, wherein the target object includes plane pattern.
5. a kind of image-recognizing method, which is characterized in that application is in the client, comprising:
When receiving the image recognition instruction for image to be detected, at least one model data file is loaded into memory
In；Wherein, the model data file to be formed by being trained according to individual training image to initial machine learning model；
Image recognition is carried out to described image to be detected respectively based on each model data file, obtains described image to be detected
Image recognition result.
6. according to the method described in claim 5, it is characterized in that, instructing it receiving the image recognition for target image
Before, further includes:
By model data coffret, the model data file that server issues is received.
7. a kind of model data file generating means characterized by comprising
Single sample training module, for being trained according to individual training image to initial machine learning model, it is described individual
It is white space in addition to target object in training image；
Model data file generation module is sent for what will be trained with the matched model data file of the target object
To client, so that client is based on the model data file and identifies the target object.
8. a kind of pattern recognition device, which is characterized in that be configured in client, comprising:
Model data file loading module will at least one for when receiving the image recognition instruction for image to be detected
A model data file is loaded into memory；Wherein, the model data file by according to individual training image to initial
Machine learning model is trained to be formed；
Picture recognition module, for carrying out image recognition to described image to be detected respectively based on each model data file,
Obtain the image recognition result of described image to be detected.
9. a kind of electronic equipment characterized by comprising
One or more processors；
Memory, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
Now the model data file generation method as described in claim 1-4 is any or the image as described in claim 5-6 is any are known
Other method.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
Model data file generation method as described in claim 1-4 is any is realized when execution or as described in claim 5-6 is any
Image-recognizing method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910599358.3A CN110334763B (en) | 2019-07-04 | 2019-07-04 | Model data file generation method, model data file generation device, model data file identification device, model data file generation apparatus, model data file identification apparatus, and model data file identification medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910599358.3A CN110334763B (en) | 2019-07-04 | 2019-07-04 | Model data file generation method, model data file generation device, model data file identification device, model data file generation apparatus, model data file identification apparatus, and model data file identification medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110334763A true CN110334763A (en) | 2019-10-15 |
| CN110334763B CN110334763B (en) | 2021-07-23 |
Family
ID=68144175
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910599358.3A Active CN110334763B (en) | 2019-07-04 | 2019-07-04 | Model data file generation method, model data file generation device, model data file identification device, model data file generation apparatus, model data file identification apparatus, and model data file identification medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110334763B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113536023A (en) * | 2021-08-06 | 2021-10-22 | 广东电网有限责任公司 | Drawing file management method and system based on BIM and Internet of things |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102637251A (en) * | 2012-03-20 | 2012-08-15 | 华中科技大学 | Face recognition method based on reference features |
| CN103294813A (en) * | 2013-06-07 | 2013-09-11 | 北京捷成世纪科技股份有限公司 | Sensitive image search method and device |
| CN106022317A (en) * | 2016-06-27 | 2016-10-12 | 北京小米移动软件有限公司 | Face identification method and apparatus |
| CN106529583A (en) * | 2016-11-01 | 2017-03-22 | 哈尔滨工程大学 | Bag-of-visual-word-model-based indoor scene cognitive method |
| CN106777350A (en) * | 2017-01-16 | 2017-05-31 | 青岛海信网络科技股份有限公司 | It is a kind of based on bayonet socket data scheming to search drawing method and device |
| CN107122375A (en) * | 2016-12-12 | 2017-09-01 | 南京理工大学 | The recognition methods of image subject based on characteristics of image |
| CN107527058A (en) * | 2017-07-25 | 2017-12-29 | 北京理工大学 | A kind of image search method based on weighting local feature Aggregation Descriptor |
| CN108304936A (en) * | 2017-07-12 | 2018-07-20 | 腾讯科技（深圳）有限公司 | Machine learning model training method and device, facial expression image sorting technique and device |
| CN108416326A (en) * | 2018-03-27 | 2018-08-17 | 百度在线网络技术（北京）有限公司 | Face identification method and device |
| CN108491543A (en) * | 2018-03-30 | 2018-09-04 | 广州汇智通信技术有限公司 | Image search method, image storage method and image indexing system |
| CN108664921A (en) * | 2018-05-10 | 2018-10-16 | 江苏大学 | Image-recognizing method and system based on bag of words under a kind of Android platform |
| CN108921929A (en) * | 2018-06-26 | 2018-11-30 | 开放智能机器（上海）有限公司 | A kind of recognition methods of identifying system and training method and individual monocular image |
| CN109117879A (en) * | 2018-08-03 | 2019-01-01 | 南京旷云科技有限公司 | Image classification method, apparatus and system |
| CN109614508A (en) * | 2018-12-12 | 2019-04-12 | 杭州知衣科技有限公司 | A kind of image of clothing searching method based on deep learning |
| CN109948707A (en) * | 2019-03-20 | 2019-06-28 | 腾讯科技（深圳）有限公司 | Model training method, device, terminal and storage medium |
-
2019
- 2019-07-04 CN CN201910599358.3A patent/CN110334763B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102637251A (en) * | 2012-03-20 | 2012-08-15 | 华中科技大学 | Face recognition method based on reference features |
| CN103294813A (en) * | 2013-06-07 | 2013-09-11 | 北京捷成世纪科技股份有限公司 | Sensitive image search method and device |
| CN106022317A (en) * | 2016-06-27 | 2016-10-12 | 北京小米移动软件有限公司 | Face identification method and apparatus |
| CN106529583A (en) * | 2016-11-01 | 2017-03-22 | 哈尔滨工程大学 | Bag-of-visual-word-model-based indoor scene cognitive method |
| CN107122375A (en) * | 2016-12-12 | 2017-09-01 | 南京理工大学 | The recognition methods of image subject based on characteristics of image |
| CN106777350A (en) * | 2017-01-16 | 2017-05-31 | 青岛海信网络科技股份有限公司 | It is a kind of based on bayonet socket data scheming to search drawing method and device |
| CN108304936A (en) * | 2017-07-12 | 2018-07-20 | 腾讯科技（深圳）有限公司 | Machine learning model training method and device, facial expression image sorting technique and device |
| CN107527058A (en) * | 2017-07-25 | 2017-12-29 | 北京理工大学 | A kind of image search method based on weighting local feature Aggregation Descriptor |
| CN108416326A (en) * | 2018-03-27 | 2018-08-17 | 百度在线网络技术（北京）有限公司 | Face identification method and device |
| CN108491543A (en) * | 2018-03-30 | 2018-09-04 | 广州汇智通信技术有限公司 | Image search method, image storage method and image indexing system |
| CN108664921A (en) * | 2018-05-10 | 2018-10-16 | 江苏大学 | Image-recognizing method and system based on bag of words under a kind of Android platform |
| CN108921929A (en) * | 2018-06-26 | 2018-11-30 | 开放智能机器（上海）有限公司 | A kind of recognition methods of identifying system and training method and individual monocular image |
| CN109117879A (en) * | 2018-08-03 | 2019-01-01 | 南京旷云科技有限公司 | Image classification method, apparatus and system |
| CN109614508A (en) * | 2018-12-12 | 2019-04-12 | 杭州知衣科技有限公司 | A kind of image of clothing searching method based on deep learning |
| CN109948707A (en) * | 2019-03-20 | 2019-06-28 | 腾讯科技（深圳）有限公司 | Model training method, device, terminal and storage medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113536023A (en) * | 2021-08-06 | 2021-10-22 | 广东电网有限责任公司 | Drawing file management method and system based on BIM and Internet of things |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110334763B (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111368685B (en) | Method and device for identifying key points, readable medium and electronic equipment | |
| CN110288049A (en) | Method and apparatus for generating image recognition model | |
| JP2023547917A (en) | Image segmentation method, device, equipment and storage medium | |
| CN110110811A (en) | Method and apparatus for training pattern, the method and apparatus for predictive information | |
| CN110298413A (en) | Image characteristic extracting method, device, storage medium and electronic equipment | |
| CN110162670A (en) | Method and apparatus for generating expression packet | |
| CN109919244A (en) | Method and apparatus for generating scene Recognition model | |
| CN109829432A (en) | Method and apparatus for generating information | |
| CN110413812A (en) | Training method, device, electronic equipment and the storage medium of neural network model | |
| CN109977839A (en) | Information processing method and device | |
| CN109815365A (en) | Method and apparatus for handling video | |
| CN109800730A (en) | The method and apparatus for generating model for generating head portrait | |
| CN109947989A (en) | Method and apparatus for handling video | |
| CN109934142A (en) | Method and apparatus for generating the feature vector of video | |
| CN113468344A (en) | Entity relationship extraction method and device, electronic equipment and computer readable medium | |
| CN110097004B (en) | Facial expression recognition method and device | |
| CN110288532B (en) | Method, apparatus, device and computer readable storage medium for generating whole body image | |
| CN110334650A (en) | Object detecting method, device, electronic equipment and storage medium | |
| CN110414450A (en) | Keyword detection method, apparatus, storage medium and electronic equipment | |
| CN109829431A (en) | Method and apparatus for generating information | |
| CN110110693A (en) | The method and apparatus of face character for identification | |
| CN110362698A (en) | A kind of pictorial information generation method, device, mobile terminal and storage medium | |
| CN110008926A (en) | The method and apparatus at age for identification | |
| CN110334763A (en) | Model data file generation, image-recognizing method, device, equipment and medium | |
| WO2023143118A1 (en) | Image processing method and apparatus, device, and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |