CN110377218A - Data processing method, device, computer equipment and storage medium - Google Patents
Data processing method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110377218A CN110377218A CN201910562432.4A CN201910562432A CN110377218A CN 110377218 A CN110377218 A CN 110377218A CN 201910562432 A CN201910562432 A CN 201910562432A CN 110377218 A CN110377218 A CN 110377218A
- Authority
- CN
- China
- Prior art keywords
- file
- target area
- model
- trained
- image recognition
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0481—Interaction techniques based on graphical user interfaces [GUI] based on specific properties of the displayed interaction object or a metaphor-based environment, e.g. interaction with desktop elements like windows or icons, or assisted by a cursor's changing behaviour or appearance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0487—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
- G06F3/0488—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/16—Sound input; Sound output
- G06F3/167—Audio in a user interface, e.g. using voice commands for navigating, audio feedback
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
This application involves a kind of data processing method, device, computer equipment and storage mediums.The described method includes: displaying interface and the corresponding touch operation of acquisition current file, target area is determined from current presentation interface according to the location information of touch operation, by the corresponding image recognition model trained of target area input current file, target object in identification object region, obtain audio data corresponding with target object, playing audio-fequency data.Region division is carried out by the touch operation at the displaying interface in real time to file, determine the region where target object, and identify target object, obtain audio data corresponding with target object, playing audio-fequency data passes through automatic identification target object, it is only necessary to establish linking for file and model, and each target object and audio data link, so that data processing becomes highly efficient.
Description
Technical field
This application involves field of computer technology more particularly to a kind of data processing method, device, computer equipment and deposit
Storage media.
Background technique
Existing interface alternation, is substantially by touching screen operator, triggers corresponding control according to touch operation, obtains
Audio data corresponding with control is taken, playing audio-fequency data directly simply links audio data, directly passes through link
Corresponding data are obtained, need to carry out data link design to the control in each displayed page, data-handling efficiency is low.
Summary of the invention
In order to solve the above-mentioned technical problem, it this application provides a kind of data processing method, device, computer equipment and deposits
Storage media.
In a first aspect, this application provides a kind of data processing methods, comprising:
The displaying interface of acquisition current file and corresponding touch operation；
Target area is determined from current presentation interface according to the location information of touch operation；
Target area is inputted into the corresponding image recognition model trained of current file, the target in identification object region
Object；
Obtain audio data corresponding with target object, playing audio-fequency data.
Second aspect, this application provides a kind of data processing equipments, comprising:
Data acquisition module, for obtain current file displaying interface and corresponding touch operation；
Target area determining module determines target area for the location information according to touch operation from current presentation interface
Domain；
The input current file corresponding image recognition model trained in target area is identified target area by identification module
Target object in domain；
Playing module, for obtaining audio data corresponding with target object, playing audio-fequency data.
A kind of computer equipment can be run on a memory and on a processor including memory, processor and storage
Computer program, the processor perform the steps of when executing the computer program
The displaying interface of acquisition current file and corresponding touch operation；
Target area is determined from current presentation interface according to the location information of touch operation；
Target area is inputted into the corresponding image recognition model trained of current file, the target in identification object region
Object；
Obtain audio data corresponding with target object, playing audio-fequency data.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
It is performed the steps of when row
The displaying interface of acquisition current file and corresponding touch operation；
Target area is determined from current presentation interface according to the location information of touch operation；
Target area is inputted into the corresponding image recognition model trained of current file, the target in identification object region
Object；
Obtain audio data corresponding with target object, playing audio-fequency data.
Above-mentioned data processing method, device, computer equipment and storage medium, which comprises obtain current file
Displaying interface and corresponding touch operation, according to the location information of touch operation from current presentation interface determine target area
Target area is inputted the corresponding image recognition model trained of current file by domain, the target object in identification object region,
Obtain audio data corresponding with target object, playing audio-fequency data.Pass through the touch operation at the displaying interface in real time to file
Region division is carried out, determines the region where target object, and identify target object, obtains audio number corresponding with target object
According to playing audio-fequency data passes through automatic identification target object, it is only necessary to establish file and model link and each target
Object is linked with audio data, so that data processing becomes highly efficient.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows and meets implementation of the invention
Example, and be used to explain the principle of the present invention together with specification.
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, for those of ordinary skill in the art
Speech, without any creative labor, is also possible to obtain other drawings based on these drawings.
Fig. 1 is the applied environment figure of data processing method in one embodiment；
Fig. 2 is the flow diagram of data processing method in one embodiment；
Fig. 3 is the structural block diagram of data processing equipment in one embodiment；
Fig. 4 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the application, instead of all the embodiments.Based on the embodiment in the application, ordinary skill people
Member's every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
Fig. 1 is the applied environment figure of data processing method in one embodiment.Referring to Fig.1, the data processing method application
In data processing system.The data processing system includes terminal 110 and server 120.Terminal 110 and server 120 pass through net
Network connection.Terminal 110 obtain current file displaying interface and corresponding touch operation, according to the location information of touch operation from
Target area is determined in current presentation interface, and target area is inputted into the corresponding image recognition model trained of current file,
Target object in identification object region obtains audio data corresponding with target object, playing audio-fequency data.Wherein, terminal
110 specifically can be terminal console or mobile terminal, and mobile terminal specifically can be in mobile phone, tablet computer, laptop etc.
At least one.Server 120 can be realized with the server cluster of the either multiple server compositions of independent server.
As shown in Fig. 2, in one embodiment, providing a kind of data processing method.The present embodiment is mainly in this way
It is illustrated applied to the terminal 110 (or server 120) in above-mentioned Fig. 1.Referring to Fig. 2, which is specifically wrapped
Include following steps:
Step S201, obtain current file displaying interface and corresponding touch operation.
Specifically, current file refers to that the file that can be played out in terminal includes video file and document.Show boundary
Face be document terminal displayed page and video file terminal plays video frame.Touch operation is that user is showing interface
The operation of execution, touch operation can be click, sliding etc. for executing the action event interacted with terminal interface.
In one embodiment, before step S201, further includes: the image recognition mould binding each file and having trained
Type saves the corresponding relationship of each file with the image recognition model trained.
Step S202 determines target area according to the location information of touch operation from current presentation interface.
Specifically, the location information of touch operation is that touch operation is showing the touch location on interface, location information packet
Include at least one set of coordinate data.Target area is the region comprising target object in current presentation interface, and target area can be with
Partial region or whole region for current presentation interface, wherein target object can for true people, animal, water fruits and vegetables,
Flowers and plants and trees can also be cartoon character, and wherein cartoon character is the image etc. in common cartoon and animation.
In one embodiment, step S202, comprising: obtain preset window, preset window includes window size, according to position
Confidence, which ceases, determines that preset window is showing the region in interface, using the corresponding region of preset window as target area.
Specifically, preset window is preconfigured window, and the size user of the size including window, window can make by oneself
Justice is also possible to the window that server defines, and wherein in same file, the region area of each target area can be identical
It can not also be identical.As can the region area according to shared by the maximum target object in each file define target area face
Product, the window size that can also pre-define the corresponding target area of each touch location of storage define target area size.Such as
The size that preset window can be set in same file is A*B, and wherein A*B is maximum according to the area accounting in file
What target object determined.Also the size that the corresponding preset window of position A1 in displayed page A can be set is C*D, position A2
The size of corresponding preset window is E*F.According to location information determine preset window show interface in target area when
Specific method of determination customized can be arranged, such as can be using location information as its of the center of preset window or target area
In a vertex etc., target area is determined according to location information and preset window.
In one embodiment, show to include input window information preset control in interface, before obtaining preset window, also
It include: the window size information for receiving user and being inputted in preset control, according to window size Automatic generation of information preset window.
Specifically, preset control is the control for receiving the customized window information of user's input.Preset window is root
According to the window that user is automatically generated by the customized window information of preset control, wherein window information includes the window ruler of window
It is very little, such as the length and width of window, preset window is generated according to length and width information.If user is defined in the pre-set space
Length of window is A, width B.After generating preset window, user also changes the dimension information of preset window, passes through sliding window
With change preset window dimension information so that target object is in the corresponding region of preset window.
In one embodiment, the corresponding location information of touch operation of user is detected, wherein location information can be one
A touch operation determination, it can also be through multiple touch operations determination.If user is in four of displayed page of document
It is clicked four times on different positions, obtains the coordinate data of four positions, determine a region according to four coordinate datas.
In one embodiment, step S202, comprising: determine that slide is corresponding according to the location information of slide
Enclosed region, using enclosed region as target area.
Specifically, slide includes at least one operation, and there may be do not advise in the region delimited after execution slide
Then or inc situation pre-processes the region of delimitation, obtains when there is irregular or inc situation
The rectangular area of the rule of one closure, using the rectangular area of the rule of closure as target area.
The input current file corresponding image recognition model trained in target area is identified target area by step S203
Target object in domain.
Specifically, the image recognition model trained refers to is instructed by the image for largely carrying target object label
The image recognition model obtained after white silk.When being trained to iconic model, an iconic model can identify multiple targets pair
As, such as identification cat, dog, rabbit animal, wherein animal includes the image and Ka Te zoomorphism that real animals are shot.
In one embodiment, before step S203, further includes: the step of generating the image recognition model trained,
Specific steps include: building initial pictures identification model, obtain the training set comprising multiple training images, each training image
Comprising object to be identified and corresponding label, each training image is inputted into initial pictures identification model, obtains each training figure
Whether the recognition result of picture judges initial pictures identification model according to the recognition result of each training image and corresponding label
Meet the model condition of convergence, when initial image recognition model meets the model condition of convergence, the image recognition mould trained
Type updates the parameter of initial pictures identification model when initial image recognition model does not meet the model condition of convergence, until initial
Image recognition model meets the model condition of convergence, the image recognition model trained.
Specifically, initial pictures identification model constructs in advance, for identification the mathematics of the object to be identified in image
Model.Mathematical model is common deep learning network model, neural network model etc..Training image includes but is not limited to clap
Take the photograph image, the designed image of designer etc. of equipment acquisition.Label is used to identify the label data of object to be identified, number of tags
According to can be customized, such as correspond to different objects to be identified using different mode bits, the object that can be identified such as model includes 3
Kind, then it include 3 mode bits, as being ordered as { cat, dog, rabbit } after defining, then { 1,0,0 } indicates cat, and { 0,1,0 } indicates
Dog, then { 0,0,1 } indicates rabbit.It can also directly be demarcated with number, such as number 1 indicates that cat, number 2 indicate dog, number 3
Indicate rabbit.
Each training image is inputted into initial pictures identification model, by initial pictures identification model to each training image
Identified, obtain the corresponding recognition result of each training image, judge each training image recognition result whether with label
Matching, matching indicate to be identified as function, unsuccessfully indicate to identify mistake.The condition of convergence is suitable for the convergence state of judgment models, receives
The state of holding back can be to be determined by the recognition accuracy of model, can also be determined according to the penalty values of model, wherein losing
The function of value can be customized, can such as define the formula for the variance for calculating characteristics of image as loss function, can also define
The formula of the logarithm of the variance of characteristics of image is calculated as loss function.Initial pictures identification model when by penalty values being minimum
As the image recognition model trained.
Initial pictures identification model does not meet the condition of convergence, when model is not converged, the parameter of more new model, and the parameter of model
Using common model update method when update, model parameter is such as updated using minimal gradient descent method.
In one embodiment, using the frequency of training of model as the model condition of convergence, such as when the frequency of training of model reaches
When to 10000 times, by the initial pictures identification model after training 1000 times as the image recognition model trained.When model
When frequency of training is not up to 10000 times, continue training pattern, model parameter is updated, until the frequency of training of model reaches 1000
It is secondary.
In one embodiment, the recognition result of each training image and the matching result of corresponding label are counted, is obtained
Whether correct recognition rata, correct judgment discrimination are greater than default correct recognition rata, preset just when correct recognition rata is greater than or equal to
True discrimination, the image recognition model trained are first into updating when correct recognition rata is less than default correct recognition rata
The parameter of beginning image recognition model.
Specifically, correct recognition rata is after initial pictures identification model identifies training image, to count recognition result
Obtained discrimination, recognition result match with corresponding label, and identification is correct, and recognition result is mismatched with corresponding label
, identify that mistake obtains correct recognition rata, it is pre- whether correct judgment discrimination is greater than by mistake of statistics rate and/or accuracy
The default correct recognition rata being first arranged, when being greater than default correct recognition rata, model convergence, the image recognition mould trained
Type updates model parameter conversely, model is not converged, until model is restrained, the image recognition model trained.
In one embodiment, the image recognition model trained is TensorFlow deep learning model.
In one embodiment, cloud configuration file and local profile are obtained, wherein configuration file includes each file
The image recognition model corresponding relationship trained, when the identical file in cloud configuration file and local profile is corresponding
When the image recognition model trained is not identical, using identical file, the corresponding image trained is known in cloud configuration file
Other model, replacement corresponding image recognition model trained in local profile.
Specifically, cloud configuration file and local profile are for storage file and the image recognition model trained
Corresponding relationship file.It can be used to search each file downloaded in terminal and the corresponding figure trained in local profile
As identification model, cloud configuration file is the text of each file stored in server and the corresponding image recognition model trained
Part.Cloud configuration file in server is updated as the data in server update, the update in local profile with
The update of data in terminal and update.It is corresponding with file each in local profile when cloud configuration file updates
Whether trained image recognition model is updated, and is updated if it exists, then it is corresponding to download identical file in cloud configuration file
The image recognition model trained, using the original image trained of the image recognition model replacement of having trained after downloading
Identification model, and update the corresponding relationship of the corresponding image recognition model trained of this document in local profile.Wherein
Judging whether cloud configuration file updates can be used to identify not by generation time, the build version number etc. for judging configuration file
Whether different with the information of cloud configuration file, when the generation time or build version number etc. is different, cloud configuration file has been carried out more
Newly.Similarly may determine that the model in cloud configuration file whether update.Image recognition model is updated, so that recognition result more accords with
It closes user demand or recognition accuracy is higher.
In one embodiment, image recognition model is updated, when introducing new training data, the image trained is known
Other model is further trained, and the updated image recognition model trained is obtained.Update the image recognition trained
Model can more accurately identify image, to promote user experience.
Step S204 obtains audio data corresponding with target object, playing audio-fequency data.
Specifically, audio data is the audio data corresponding with each target object being stored in advance in the server,
Sound intermediate frequency data can be customized, such as Chinese including target object, translator of English, target object explain in detail
Deng.If target object is a doggie, then the audio data that can be returned includes Chinese " dog ", and English name " dog " is named
At least one of sound " tearful " and related introduction.
Above-mentioned data processing method, comprising: obtain current file displaying interface and corresponding touch operation, according to touch
The location information of operation determines target area from current presentation interface；By target area input, current file is corresponding has trained
Image recognition model, the target object in identification object region obtains audio data corresponding with target object, broadcasting audio
Data.The target object of the file in interface is shown by real-time identification terminal, is obtained the corresponding audio data of target in real time, is subtracted
The process linked to each target object is lacked, when the frequency that each target object in file occurs is higher and very fast
When, link process is complicated, and the application only needs the image for having trained all or part of region input shown in interface to know
Other model can determine corresponding audio data according to recognition result, and the mapping relations of data are more simple, reduce data
Data-handling efficiency is improved while memory size.
In a specific embodiment, above-mentioned data processing method, comprising:
For convenience of explanation, the image recognition model trained in the present embodiment is TensorFlow deep learning mould
Type, file are to draw this.
The various TensorFlow Lite models trained are generated in server end, can be used for identifying animal, plant etc.
Etc. the corresponding image recognition model trained of targets.During model training, the various cartoons drawn in this can also be added
Image, to increase the accuracy rate of model identification.To training pattern, it can be periodically trained update, some by new production draw
Training set is added in this cartoon character, and the accuracy rate for identifying model is constantly improve.
In one embodiment, model training is carried out in server-side with a large amount of pictures, for example trains an animal model, used
The picture of various animals, true and cartoon animal picture, is trained.Need to prepare a training set, such as cat
Various pictures are placed in a file, and this document folder is named as cat, when training, it is intended to model learning to this document
All picture patterns under folder are cat, in addition build a file again, are named as dog, principle is same.Such pictures, with now
The model for having very mature T ensorFlow training to obtain can identify the species drawn in this.
This when is drawn in production, the TensorFlow model trained that may be used in this will be drawn and bound with drawing this,
Such as the race between tortoise and rabbit draws this, and it is related to the hare and the tortoise, it will identify rabbit, identify tortoise or other relevant animals
The TensorFlow model trained is drawn this with this and is bound, and model ID is recorded and is drawn in this information, configuration file is obtained.
In one embodiment, can be by the corresponding training pattern of a major class species for the training of model, for example move
Species correspond to a model, the corresponding model of plant, the corresponding model of Building class, if one is drawn in this comprising animal
And plant, just the corresponding model of animal and plant is bound with this is drawn.
When this APP is drawn in user's use, selects any one to draw this, download this and draw this while, downloading draws this binding with this
The TensorFlow model trained, when clicking some object, the available TensorFlow mould trained downloaded of APP
Type carries out image recognition, after recognizing the specific ID of object or title, service end interface is called to obtain pertinent audio information, audio
I.e. comprising explaining in detail to the object in information, APP carries out audio broadcasting.
In one embodiment, user clicks picture, and procedure identification, will be current to the coordinate points and current page clicked
Page and coordinate points are transmitted to recognizer, and recognizer can be identified according to TensorFlow model all in current image to be known
The coordinate range of other object searches the identification object that the range of click coordinate point is included, and identification object is finally returned to journey
Sequence.
In one embodiment, it when model modification, does not need to update APP, model timing updates, and new draws this downloading when having
When, then what is downloaded is updated model, and for the model bound in this of drawing downloaded before, client can be supported actively more
The function of new model.
In one embodiment, more new capital is consistency operation, when there is new drawing originally to produce, by card therein
The picture for leading to image is added in training set picture library, is trained again to model, with the accuracy of lift scheme.When client again
When secondary downloading model, what be will acquire is the higher training pattern of accuracy.In view of if each newly draws this production
If carrying out model training again, compare consuming resource, it is therefore proposed that timing updates, for example updates once for one month, certainly,
Specific renewal time can determine according to demand.The pictures for intercepting out in this are drawn to all new productions in renewal time section
It is added in picture library, is uniformly trained.
In actual use, since part cartoon character is relatively more abstract, recognition result accuracy rate is caused to decline, in order to
Recognition accuracy is improved, needs timely to update model.Such as there are various transformations to the image of cat, recognizer is for object
Kind identification place one's entire reliance upon the coverage of training set, for example for the training set of cat, we contain the picture of true cat,
The picture of hellokitty goes identification Garfield with the model that such training set trains, and possible accuracy can be declined slightly,
If drawing this when in production Garfield, the cartoon character of Garfield is added in training set, is trained again, then it can be with lift scheme
Recognition accuracy.
Fig. 2 is the flow diagram of data processing method in one embodiment.Although should be understood that the process of Fig. 2
Each step in figure is successively shown according to the instruction of arrow, but these steps are not the inevitable sequence indicated according to arrow
Successively execute.Unless expressly stating otherwise herein, there is no stringent sequences to limit for the execution of these steps, these steps can
To execute in other order.Moreover, at least part step in Fig. 2 may include multiple sub-steps or multiple stages,
These sub-steps or stage are not necessarily to execute completion in synchronization, but can execute at different times, these
Sub-step perhaps the stage execution sequence be also not necessarily successively carry out but can be with the son of other steps or other steps
Step or at least part in stage execute in turn or alternately.
In one embodiment, as shown in figure 3, providing a kind of data processing equipment, comprising:
Data acquisition module 201, for obtain current file displaying interface and corresponding touch operation.
Target area determining module 202 determines mesh for the location information according to touch operation from current presentation interface
Mark region.
Identification module 203, for identifying the corresponding image recognition model trained of target area input current file
Target object in target area.
Playing module 204, for obtaining audio data corresponding with target object, playing audio-fequency data.
In one embodiment, above-mentioned data processing equipment, further includes: model generation module, wherein model generation module,
Include:
Model construction unit, for constructing initial pictures identification model；
Training data acquiring unit includes that training for multiple training images is gathered for obtaining, and each training image includes
Object to be identified and corresponding label；
Recognition unit obtains the knowledge of each training image for each training image to be inputted initial pictures identification model
Other result；
Model restrains judging unit, and for the recognition result and corresponding label according to each training image, judgement is initial
Whether image recognition model meets the model condition of convergence；
Model generation unit, the figure for having been trained when initial image recognition model meets the model condition of convergence
As identification model, when initial image recognition model does not meet the model condition of convergence, the parameter of initial pictures identification model is updated,
Until whether initial pictures identification model meets the model condition of convergence, the image recognition model trained.
In one embodiment, above-mentioned model convergence judging unit be also used to count each training image recognition result and
The matching result of corresponding label, obtains correct recognition rata, and whether correct judgment discrimination is greater than default correct recognition rata.
Model generation unit is also used to be greater than or equal to default correct recognition rata, the figure trained when correct recognition rata
As identification model, when correct recognition rata is less than default correct recognition rata, into the parameter for updating initial pictures identification model.
In one embodiment, above-mentioned target area determining module is specifically used for obtaining preset window, and preset window includes
Window size determines preset window according to location information, the region in shown interface, and the corresponding region of preset window is made
For target area.
In one embodiment, target area determining module is also used to receive the window ruler that user inputs in preset control
Very little information, according to window size Automatic generation of information preset window, wherein showing default comprising input window information in interface
Control.
In one embodiment, above-mentioned target area determining module is also used to be determined according to the location information of slide and slide
It is dynamic to operate corresponding enclosed region, using enclosed region as target area.
In one embodiment, above-mentioned data processing equipment, further includes:
Configuration file obtains module, and for obtaining cloud configuration file and local profile, wherein configuration file includes each
A file and the image recognition model corresponding relationship trained；
Model modification module, for having been trained when the identical file in cloud configuration file and local profile is corresponding
Image recognition model it is not identical when, using identical file in cloud configuration file the corresponding image recognition mould trained
Type, replacement corresponding image recognition model trained in local profile.
Fig. 4 shows the internal structure chart of computer equipment in one embodiment.The computer equipment specifically can be Fig. 1
In terminal 110 (or server 120).As shown in figure 4, it includes total by system that the computer equipment, which includes the computer equipment,
Processor, memory, network interface, input unit and the display screen of line connection.Wherein, memory includes that non-volatile memories are situated between
Matter and built-in storage.The non-volatile memory medium of the computer equipment is stored with operating system, can also be stored with computer journey
Sequence when the computer program is executed by processor, may make processor to realize data processing method.It can also be stored up in the built-in storage
There is computer program, when which is executed by processor, may make processor configuration for executing data processing.Computer
The display screen of equipment can be liquid crystal display or electric ink display screen, and the input unit of computer equipment can be display
The touch layer covered on screen is also possible to the key being arranged on computer equipment shell, trace ball or Trackpad, can also be outer
Keyboard, Trackpad or mouse for connecing etc..
It will be understood by those skilled in the art that structure shown in Fig. 4, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, data processing equipment provided by the present application can be implemented as a kind of shape of computer program
Formula, computer program can be run in computer equipment as shown in Figure 4.Composition can be stored in the memory of computer equipment should
Each program module of data processing equipment, for example, data acquisition module shown in Fig. 3 201, target area determining module
202A, identification module 203 and playing module 204.The computer program that each program module is constituted makes processor execute this theory
Step in the data processing method of each embodiment of the application described in bright book.
For example, computer equipment shown in Fig. 4 can pass through the data acquisition mould in data processing equipment as shown in Figure 3
Block 201 executes the displaying interface for obtaining current file and corresponding touch operation.Computer equipment can be true by target area
Cover half block 202 is executed determines target area according to the location information of touch operation from current presentation interface.Computer equipment can
Target area is inputted into the corresponding image recognition model trained of current file to execute by identification module 203, identifies mesh
Mark the target object in region.Computer equipment can be executed by playing module 204 and obtain audio corresponding with target object
Data, playing audio-fequency data.
In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, processor performs the steps of acquisition when executing computer program ought be above
The displaying interface of part and corresponding touch operation；Target area is determined from current presentation interface according to the location information of touch operation
Domain；Target area is inputted into the corresponding image recognition model trained of current file, the target object in identification object region；
Obtain audio data corresponding with target object, playing audio-fequency data.
In one embodiment, the image for generating and having trained also is performed the steps of when processor executes computer program
Identification model method includes: building initial pictures identification model；Obtain the training set comprising multiple training images, each training
Image includes object to be identified and corresponding label；Each training image is inputted into initial pictures identification model, obtains each instruction
Practice the recognition result of image；According to the recognition result of each training image and corresponding label, initial pictures identification model is judged
Whether the model condition of convergence is met；When initial image recognition model meets the model condition of convergence, the image trained is known
Other model；When initial image recognition model does not meet the model condition of convergence, the parameter of initial pictures identification model is updated, until
Whether initial pictures identification model meets the model condition of convergence, the image recognition model trained.
In one embodiment, the model condition of convergence includes default correct recognition rata, according to the identification of each training image
As a result with corresponding label, judge whether initial pictures identification model meets the model condition of convergence, comprising: each training figure of statistics
The matching result of the recognition result of picture and corresponding label, obtains correct recognition rata；It is default whether correct judgment discrimination is greater than
Correct recognition rata；When correct recognition rata is greater than or equal to default correct recognition rata, the image recognition model trained；When just
When true discrimination is less than default correct recognition rata, into the parameter of initial pictures identification model.
In one embodiment, target area is determined from displaying interface according to the location information of touch operation, comprising: obtain
Preset window is taken, preset window includes window size；Preset window is determined according to location information, is showing the region in interface；
Using the corresponding region of preset window as target area.
In one embodiment, the preset control in interface comprising input window information is shown, before obtaining preset window,
Processor also performs the steps of when executing computer program receives the window size information that user inputs in preset control,
According to window size Automatic generation of information preset window.
In one embodiment, touch operation is slide, according to the location information of touch operation from displaying interface
Determine target area, comprising: the corresponding enclosed region of slide delimited according to the location information of slide；By enclosed region
As target area.
In one embodiment, processor execute computer program when also perform the steps of obtain cloud configuration file and
Local profile, wherein configuration file includes each file and the image recognition model corresponding relationship trained；When cloud configures
When the image recognition model that identical file in file and local profile is corresponding to have trained is not identical, use is identical
File corresponding image recognition model trained in cloud configuration file replaces corresponding in local profile trained
Image recognition model.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of the displaying interface for obtaining current file and corresponding touch operation when being executed by processor；According to
The location information of touch operation determines target area from current presentation interface；Target area input current file is corresponding
Trained image recognition model, the target object in identification object region；Audio data corresponding with target object is obtained, is played
Audio data.
In one embodiment, the figure for generating and having trained also is performed the steps of when computer program is executed by processor
As identification model method includes: building initial pictures identification model；Obtain the training set comprising multiple training images, Ge Gexun
Practicing image includes object to be identified and corresponding label；Each training image is inputted into initial pictures identification model, is obtained each
The recognition result of training image；According to the recognition result of each training image and corresponding label, judge that initial pictures identify mould
Whether type meets the model condition of convergence；When initial image recognition model meets the model condition of convergence, the image trained
Identification model；When initial image recognition model does not meet the model condition of convergence, the parameter of initial pictures identification model is updated, directly
Whether meet the model condition of convergence to initial pictures identification model, the image recognition model trained.
In one embodiment, the model condition of convergence includes default correct recognition rata, according to the identification of each training image
As a result with corresponding label, judge whether initial pictures identification model meets the model condition of convergence, comprising: each training figure of statistics
The matching result of the recognition result of picture and corresponding label, obtains correct recognition rata；It is default whether correct judgment discrimination is greater than
Correct recognition rata；When correct recognition rata is greater than or equal to default correct recognition rata, the image recognition model trained；When just
When true discrimination is less than default correct recognition rata, into the parameter of initial pictures identification model.
In one embodiment, target area is determined from displaying interface according to the location information of touch operation, comprising: obtain
Preset window is taken, preset window includes window size；Preset window is determined according to location information, is showing the region in interface；
Using the corresponding region of preset window as target area.
In one embodiment, the preset control in interface comprising input window information is shown, before obtaining preset window,
It is also performed the steps of when computer program is executed by processor and receives the window size letter that user inputs in preset control
Breath, according to window size Automatic generation of information preset window.
In one embodiment, touch operation is slide, according to the location information of touch operation from displaying interface
Determine target area, comprising: the corresponding enclosed region of slide delimited according to the location information of slide；By enclosed region
As target area.
In one embodiment, it is also performed the steps of when computer program is executed by processor and obtains cloud configuration file
And local profile, wherein configuration file includes each file and the image recognition model corresponding relationship trained；When cloud is matched
Set the corresponding image recognition model trained of the identical file in file and local profile it is not identical when, use is identical
File in cloud configuration file the corresponding image recognition model trained, replace and corresponding in local profile instructed
Experienced image recognition model.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in a non-volatile computer and can be read
In storage medium, the program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, provided herein
Each embodiment used in any reference to memory, storage, database or other media, may each comprise non-volatile
And/or volatile memory.Nonvolatile memory may include that read-only memory (ROM), programming ROM (PROM), electricity can be compiled
Journey ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include random access memory
(RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms, such as static state RAM
(SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhanced SDRAM
(ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) directly RAM (RDRAM), straight
Connect memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
It should be noted that, in this document, the relational terms of such as " first " and " second " or the like are used merely to one
A entity or operation with another entity or operate distinguish, without necessarily requiring or implying these entities or operation it
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant are intended to
Cover non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or setting
Standby intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in the process, method, article or apparatus that includes the element.
The above is only a specific embodiment of the invention, is made skilled artisans appreciate that or realizing this hair
It is bright.Various modifications to these embodiments will be apparent to one skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and applied principle and features of novelty phase one herein
The widest scope of cause.
Claims (10)
1. a kind of data processing method, which is characterized in that the described method includes:
The displaying interface of acquisition current file and corresponding touch operation；
Target area is determined from the current presentation interface according to the location information of the touch operation；
The target area is inputted into the corresponding image recognition model trained of the current file, identifies the target area
In target object；
Obtain audio data corresponding with the target object, playing audio data.
2. the method according to claim 1, wherein the location information according to the touch operation is from described
It shows and determines target area in interface, comprising:
Preset window is obtained, the preset window includes window size；
The preset window is determined according to the positional information, the region in the displaying interface；
Using the corresponding region of the preset window as the target area.
3. according to the method described in claim 2, it is characterized in that, described show presetting comprising input window information in interface
Control, before the acquisition preset window, further includes:
Receive the window size information that user inputs in the preset control；
According to preset window described in the window size Automatic generation of information.
4. the method according to claim 1, wherein the touch operation be slide, it is described according to
The location information of touch operation determines target area from the displaying interface, comprising:
The corresponding enclosed region of the slide is determined according to the location information of the slide；
Using the enclosed region as the target area.
5. method according to claim 1 to 4, which is characterized in that the method also includes:
Cloud configuration file and local profile are obtained, wherein configuration file includes each file and the image recognition mould trained
Type corresponding relationship；
When the corresponding image recognition mould trained of the identical file in the cloud configuration file and the local profile
When type is not identical, using identical file, the corresponding image recognition model trained, replacement exist in the cloud configuration file
The corresponding image recognition model trained in the local profile.
6. a kind of data processing equipment, which is characterized in that described device includes:
Data acquisition module, for obtain current file displaying interface and corresponding touch operation；
Target area determining module determines mesh for the location information according to the touch operation from the current presentation interface
Mark region；
Identification module is known for the target area to be inputted the corresponding image recognition model trained of the current file
Target object in the not described target area；
Playing module, for obtaining audio data corresponding with the target object, playing audio data.
7. device according to claim 6, which is characterized in that the target area determining module is specifically used for obtaining default
Window, the preset window include window size, determine the preset window according to the positional information, at the displaying interface
In region, using the corresponding region of the preset window as the target area.
8. device according to claim 6, which is characterized in that described device further include:
Configuration file obtains module, and for obtaining cloud configuration file and local profile, wherein configuration file includes each text
Part and the image recognition model corresponding relationship trained；
Model modification module, it is corresponding for working as the identical file in the cloud configuration file and the local profile
When trained image recognition model is not identical, using identical file in the cloud configuration file the corresponding image trained
Identification model, replacement corresponding image recognition model trained in the local profile.
9. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor realizes any one of claims 1 to 5 institute when executing the computer program
The step of stating method.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 5 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910562432.4A CN110377218B (en) | 2019-06-26 | 2019-06-26 | Data processing method, data processing device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910562432.4A CN110377218B (en) | 2019-06-26 | 2019-06-26 | Data processing method, data processing device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110377218A true CN110377218A (en) | 2019-10-25 |
| CN110377218B CN110377218B (en) | 2021-09-28 |
Family
ID=68249427
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910562432.4A Active CN110377218B (en) | 2019-06-26 | 2019-06-26 | Data processing method, data processing device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110377218B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113365132A (en) * | 2021-05-27 | 2021-09-07 | 网易有道信息技术(江苏)有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN113538661A (en) * | 2021-07-06 | 2021-10-22 | 万翼科技有限公司 | Information display method based on building model and related device |
| CN113573096A (en) * | 2021-07-05 | 2021-10-29 | 维沃移动通信（杭州）有限公司 | Video processing method, video processing device, electronic equipment and medium |
| CN113837830A (en) * | 2021-09-13 | 2021-12-24 | 珠海格力电器股份有限公司 | Product display method, display device and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103679129A (en) * | 2012-09-21 | 2014-03-26 | 中兴通讯股份有限公司 | Method and device for identifying object in image |
| CN107967110A (en) * | 2017-11-30 | 2018-04-27 | 广东小天才科技有限公司 | Playback method, playing device, electronic equipment and computer-readable recording medium |
| CN108236785A (en) * | 2018-02-08 | 2018-07-03 | 腾讯科技（深圳）有限公司 | A kind of method and device for obtaining object information |
| CN109583514A (en) * | 2018-12-19 | 2019-04-05 | 成都西纬科技有限公司 | A kind of image processing method, device and computer storage medium |
| US20190108420A1 (en) * | 2017-05-14 | 2019-04-11 | International Business Machines Corporation | Systems and methods for identifying a target object in an image |
-
2019
- 2019-06-26 CN CN201910562432.4A patent/CN110377218B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103679129A (en) * | 2012-09-21 | 2014-03-26 | 中兴通讯股份有限公司 | Method and device for identifying object in image |
| US20190108420A1 (en) * | 2017-05-14 | 2019-04-11 | International Business Machines Corporation | Systems and methods for identifying a target object in an image |
| CN107967110A (en) * | 2017-11-30 | 2018-04-27 | 广东小天才科技有限公司 | Playback method, playing device, electronic equipment and computer-readable recording medium |
| CN108236785A (en) * | 2018-02-08 | 2018-07-03 | 腾讯科技（深圳）有限公司 | A kind of method and device for obtaining object information |
| CN109583514A (en) * | 2018-12-19 | 2019-04-05 | 成都西纬科技有限公司 | A kind of image processing method, device and computer storage medium |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113365132A (en) * | 2021-05-27 | 2021-09-07 | 网易有道信息技术(江苏)有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN113573096A (en) * | 2021-07-05 | 2021-10-29 | 维沃移动通信（杭州）有限公司 | Video processing method, video processing device, electronic equipment and medium |
| CN113538661A (en) * | 2021-07-06 | 2021-10-22 | 万翼科技有限公司 | Information display method based on building model and related device |
| CN113837830A (en) * | 2021-09-13 | 2021-12-24 | 珠海格力电器股份有限公司 | Product display method, display device and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110377218B (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110377218A (en) | Data processing method, device, computer equipment and storage medium | |
| CN110442417A (en) | Feature Extraction Method, machine learning method and its device | |
| CN108132887B (en) | User interface method of calibration, device, software testing system, terminal and medium | |
| US20150271218A1 (en) | All-Electronic Ecosystems for Design and Collaboration | |
| US20160320931A1 (en) | Career history exercise data visualization | |
| CN108197030B (en) | Software interface automatic test cloud platform device based on deep learning and test method | |
| CN108021626A (en) | Page composing method, device and equipment | |
| CN111752557A (en) | Display method and device | |
| CN106030612A (en) | Providing photo heat maps | |
| Courtiol et al. | Isoscape computation and inference of spatial origins with mixed models using the R package IsoriX | |
| CN112131837B (en) | Service report configuration method, device, computer equipment and storage medium | |
| CN109255826A (en) | Chinese training image generation method, device, computer equipment and storage medium | |
| CN110765015A (en) | Method for testing application to be tested and electronic equipment | |
| CN109640068A (en) | Information forecasting method, device, equipment and the storage medium of video frame | |
| CN114218514A (en) | Page generation method, device, equipment and storage medium | |
| EP2869195A1 (en) | Application coordination system, application coordination method, and application coordination program | |
| CN112132654A (en) | Method, device and storage medium for displaying house source information | |
| CN108959475A (en) | A kind of webpage setting method and device | |
| CN111352623B (en) | Page generation method and device | |
| CN110489379A (en) | Aircraft take a flight test synthesis display and data analysis Evaluation Platform | |
| CN109190019B (en) | User image generation method, electronic equipment and computer storage medium | |
| CN111352680A (en) | Information recommendation method and device | |
| CN110134815A (en) | Image processing method, device, computer equipment and storage medium | |
| CN109871214A (en) | Program code generation method, device, computer equipment and storage medium | |
| KR102529627B1 (en) | Coding methods and coding educational system |
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 |