CN109409423A - A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing - Google Patents
A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN109409423A CN109409423A CN201811196560.3A CN201811196560A CN109409423A CN 109409423 A CN109409423 A CN 109409423A CN 201811196560 A CN201811196560 A CN 201811196560A CN 109409423 A CN109409423 A CN 109409423A
- Authority
- CN
- China
- Prior art keywords
- plot
- target image
- image
- information
- image set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of image-recognizing method, device, terminal and readable storage medium storing program for executing, this method comprises: obtaining target image set to be identified；The target image set is input in the image recognition model that training is completed in advance, is based on described image identification model, determines at least one corresponding plot of the target image set；Export at least one described plot, the image recognition model completed in the present invention based on preparatory training, it can determine at least one corresponding plot of target image set to be identified and export, realize and help user's especially special population quick and precisely to obtain the meaning or content wanting to convey in image.
Description
Technical field
The present invention relates to computer vision recognition technology field more particularly to a kind of image-recognizing method, device, terminal and
Readable storage medium storing program for executing.
Background technique
With the development of artificial intelligence, image recognition rate is higher and higher, and public security has passed through Identification of Images and arrested multiple criminals
Guilty object greatly improves case-solving rate, and portrait recognition technology is more and more mature, and brings great convenience for society
Property.Computer Vision Recognition is one core of artificial intelligence field, is constantly subjected to the extensive of academia and industrial circle for a long time
Concern.The positioning or identification of object, object detection, vision object detection all have evolved to the height of a relative maturity.
Deep learning refer to machine the mechanism of human brain simulated by deep neural network come learn, judge, decision, and it is deep
Degree study is applied to many fields, identification, recognition of face, person's handwriting identification, automatic Pilot including voice etc., and shows
There is the realization for the image recognition for applying computer vision recognition technology in technology to be mainly based upon deep learning also to complete.Example
If AlphaGo study go is also same principle, a large amount of go case data is inputted to machine by the mankind, machine uses
Depth neural network handles image, finally carries out drawing inferences about other cases from one instance for go, makes accurate judgment.
Although the prior art can be realized the identification of portrait in image or object etc., but people can not according to image
The meaning or content for wanting to convey in image are rapidly and accurately captured, particularly with certain special populations, such as group visually impaired is deposited
In the poor crowd etc. of the crowd of Dyslexia or understandability, ability to express, can rapidly and accurately learn in image
The meaning for wanting to convey just is particularly important.
Summary of the invention
The present invention provides a kind of image-recognizing method, device, terminal and readable storage medium storing program for executing, to solve the prior art
Middle user cannot rapidly and accurately capture the problem of meaning or content conveyed are wanted in image according to image.
The present invention provides image-recognizing methods, are applied to terminal, this method comprises:
Obtain target image set to be identified；
The target image set is input in the image recognition model that training is completed in advance, is identified based on described image
Model determines at least one corresponding plot of the target image set；
Export at least one described plot.
Further, it is described be based on described image identification model, determine the target image set it is corresponding at least one
Plot includes:
Based on described image identification model, the fisrt feature information that target image includes in the target image set is determined
With second feature information, wherein the fisrt feature information includes shooting time information and shooting location information；
For each target image in the target image set, judge in the target image set with the presence or absence of with
Remaining target image of the fisrt feature information association of the target image；If so, according to the sequencing of shooting time, according to
The target image and shooting location information and second feature information in remaining described target image, generate a plot；
If not, generating a plot according to the shooting location information and second feature information of the target image.
Further, the second feature information comprises at least one of the following: person characteristic information, scenery characteristic information,
Object features information, animal character information, seasonal characteristic information and weather characteristics information.
Further, at least one plot described in the output includes:
For each plot at least one described plot, by the plot and the story feelings are generated
The corresponding output of the target image of section.
Further, at least one plot described in the output includes:
At least one described plot is exported in a manner of text segment；And/or
At least one described plot is exported in a manner of voice broadcast.
Further, if exporting at least one described plot in a manner of text segment, described in the output extremely
After a few plot, the method also includes:
Judge whether to receive user to the edit requests of text segment, wherein carrying in the edit requests to be edited
Story in the first text set and edited second text set；
If so, the first text set, which is updated to second character set, merges display.
Further, the training process of described image identification model includes:
For each sample image set in training set, the every of the corresponding handmarking of each sample image set is obtained
The corresponding plot of story of the mark and the mark of the story that sample image belongs in a sample image set；
According in the corresponding sample image set of the sample image set, the sample image set got
Sample image ownership story mark and the mark the corresponding plot of story, to described image identification model into
Row training.
The present invention provides a kind of pattern recognition devices, are applied to terminal, which includes:
Module is obtained, for obtaining target image set to be identified；
Determining module, for the target image set to be input in the image recognition model that training is completed in advance, base
In described image identification model, at least one corresponding plot of the target image set is determined；
Output module, for exporting at least one described plot.
Further, the determining module is specifically used for being based on described image identification model, determines the target image set
The fisrt feature information and second feature information that target image includes in conjunction, wherein the fisrt feature information includes shooting time
Information and shooting location information；For each target image in the target image set, the target image set is judged
In with the presence or absence of remaining target image with the fisrt feature information association of the target image；If so, according to shooting time
Sequencing is generated according to the shooting location information and second feature information in the target image and remaining described target image
One plot；If not, generating a story feelings according to the shooting location information and second feature information of the target image
Section.
Further, the second feature information comprises at least one of the following: person characteristic information, scenery characteristic information,
Object features information, animal character information, seasonal characteristic information and weather characteristics information.
Further, the output module, specifically for for each story feelings at least one described plot
The plot and the target image for generating the plot are corresponded to output by section.
Further, the output module, specifically for exporting at least one described story feelings in a manner of text segment
Section；And/or at least one described plot is exported in a manner of voice broadcast.
Further, the output module, if being also used to export at least one described story in a manner of text segment
Plot judges whether to receive user to the edit requests of text segment, wherein carrying in the edit requests to be edited
The first text set and edited second text set in story；If so, the first text set is updated to institute
It states the second character set and merges display.
Further, described device further include:
Training module, for it is corresponding to obtain each sample image set for each sample image set in training set
Handmarking each sample image set in sample image ownership story mark and the mark story it is corresponding
Plot；According to the corresponding sample image set of the sample image set, the sample image set got
The corresponding plot of story of the mark and the mark of the story of middle sample image ownership, to described image identification model
It is trained.
The present invention provides a kind of terminals, comprising: processor, communication interface, memory and communication bus, wherein processing
Device, communication interface, memory complete mutual communication by communication bus；
It is stored with computer program in the memory, when described program is executed by the processor, so that the place
Manage the step of device executes any of the above-described the method.
The present invention provides a kind of computer readable storage medium, it is stored with the computer program that can be executed by terminal,
When described program is run on the terminal, so that the step of terminal executes any of the above-described the method.
The present invention provides a kind of image-recognizing method, device, terminal and readable storage medium storing program for executing, this method comprises: obtaining
Target image set to be identified；The target image set is input in the image recognition model that training is completed in advance, base
In described image identification model, at least one corresponding plot of the target image set is determined；Output described at least one
A plot.The image recognition model completed in the present invention based on preparatory training, can determine target image set to be identified
It closes at least one corresponding plot and exports, realize and user's especially special population is helped quick and precisely to obtain image
In want convey the meaning or content.
Detailed description of the invention
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, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of schematic diagram for image recognition processes that the embodiment of the present invention 1 provides；
Fig. 2 is a kind of schematic diagram for image recognition process that the embodiment of the present invention 5 provides；
Fig. 3 is a kind of structural schematic diagram for terminal that the embodiment of the present invention 7 provides；
Fig. 4 is a kind of pattern recognition device schematic diagram provided in an embodiment of the present invention.
Specific embodiment
The meaning or content for wanting to convey in image to help user quick and precisely to obtain, the embodiment of the invention provides
A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing.
Image-recognizing method provided in an embodiment of the present invention, can be applied in terminal, which includes hardware layer, operation
Operating system layer on hardware layer, and the application layer of operation on an operating system.
The hardware layer include central processing unit (CPU, Central Processing Unit), memory management unit (MMU,
Memory Management Unit) and the hardware such as memory.
The operating system can be any one or more computer behaviour that image recognition is realized by process (Process)
Make system, for example, (SuSE) Linux OS, Unix operating system, Android operation system, iOS operating system or windows behaviour
Make system etc..
And the terminal can be the handheld devices such as smart phone, tablet computer in embodiments of the present invention, be also possible to
The terminal devices such as desktop computer, portable computer, server in the embodiment of the present invention and are not particularly limited, as long as can
There is the program of the code of the image-recognizing method in the embodiment of the present invention by log, realizes image recognition.
The executing subject of image recognition in the embodiment of the present invention can be being capable of caller in terminal or terminal
And execute the functional module of program.
To make the objectives, technical solutions, and advantages of the present invention clearer, make below in conjunction with the attached drawing present invention into one
Step ground detailed description, it is clear that described embodiment is only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
Every other embodiment, shall fall within the protection scope of the present invention.
Embodiment 1:
Fig. 1 be a kind of schematic diagram of image recognition processes provided in an embodiment of the present invention, the process the following steps are included:
S101: target image set to be identified is obtained.
Image-recognizing method provided in an embodiment of the present invention is applied to terminal, and the terminal is available to arrive target to be identified
Image collection.
The process that the terminal obtains target image set to be identified can be terminal acquisition image, by collected figure
As the target image being determined as in target image set, the target image set that the terminal is got in other equipment can be
It closes, the target image set that the target image set in the other equipment can be constituted for the other equipment acquired image,
It can be the target image set etc. pre-saved in the other equipment.
It may include one or more target image in target image set to be identified.
Terminal obtains after obtaining target image set to be identified and can be the image recognition request for receiving user, can be with
It is acquisition when receiving the target image set of other equipment transmission, is also possible to obtain at set time intervals, such as
Fruit is to obtain at set time intervals, then terminal preserves the time interval of the setting, and the time interval of the setting is any,
Such as can be 5 seconds, 10 minutes, 30 minutes etc., the time interval of the setting can be developer's setting, can be by user
According to the demand of itself and habit setting etc..
S102: the target image set is input in the image recognition model that training is completed in advance, is based on the figure
As identification model, at least one corresponding plot of the target image set is determined.
The mark of the story by belonging to according to sample image in sample image set, each sample image set is saved in terminal
Know and image recognition model that the story corresponding plot training of the mark is completed, therefore according to target image set,
Based on image recognition model, the corresponding at least one plot of the target image set is determined.
The image recognition model that the training is completed can determine image recognition model for sample data under big data analysis.
The target image for including at least one corresponding plot of the target image set and the target image set
It is related.
S103: at least one described plot is exported.
It, can be defeated by least one plot after terminal determines at least one plot based on image recognition model
Out, user is helped quick and precisely to understand the meaning or content that the target image in target image set is intended by.
When terminal exports at least one plot, it can be and exported in a manner of text or voice etc., which can
To be to be selected by user according to the use demand of itself or habit, it can be and preserve the preset way of output of default in terminal.
The image-recognizing method of the embodiment of the present invention can be applied to several scenes, is such as directed to group visually impaired, there is reading
The crowd of obstacle can be exported by way of voice, allow it will also be appreciated that the meaning in picture；For understandability,
The poor crowd of ability to express can solve the problem that it is difficult that it is started writing, realize checking for the leading meaning of image；It can certainly
It improves efficiency, user directly can be turned over image when sending out wechat circle of friends, QQ having a talk about or microblogging by the image-recognizing method
It is translated into text, the narration of story can be carried out with intelligent recognition plot, without browsing multiple images, that is, may know that image
In the content that is intended by.
The image recognition model completed in the embodiment of the present invention based on preparatory training, can determine target image to be identified
Gather at least one corresponding plot and export, realizes and user's especially special population is helped quick and precisely to obtain figure
The meaning or content conveyed are wanted as in.
Embodiment 2:
On the basis of the above embodiments, in the embodiment of the present invention, it is described be based on described image identification model, determine described in
At least one corresponding plot of target image set includes:
Based on described image identification model, the fisrt feature information that target image includes in the target image set is determined
With second feature information, wherein the fisrt feature information includes shooting time information and shooting location information；
For each target image in the target image set, judge in the target image set with the presence or absence of with
Remaining target image of the fisrt feature information association of the target image；If so, according to the sequencing of shooting time, according to
The target image and shooting location information and second feature information in remaining described target image, generate a plot；
If not, generating a plot according to the shooting location information and second feature information of the target image.
The second feature information comprises at least one of the following: person characteristic information, scenery characteristic information, object features letter
Breath, animal character information, seasonal characteristic information and weather characteristics information.
In order to realize the identification of image, terminal can be based on image recognition model, determine target image set to be identified
At least one corresponding plot.
The image recognition model can determine the fisrt feature information and second that target image includes in target image set
Characteristic information.
The fisrt feature information includes shooting time information and shooting location information, when which is specific
Between information, the shooting location information is available to the better addresses information such as country, province, city be at county level.
Usual shooting time information and shooting location information carry in the picture, therefore can directly extract the in the picture
One characteristic information.
The second feature information includes person characteristic information, scenery characteristic information, object features information, animal character letter
At least one of breath, seasonal characteristic information and weather characteristics information.The person characteristic information can pass through intelligent measurement personage
Role analyzes, specific available to information such as number, gender, crowd, clothing, expression or titles, is described by correlation
The description of word, user can also be manually entered person names in advance to realize the subsequent identification of person characteristic information, crowd tool
Body can include adult, old age, juvenile, baby etc. for age-colony.The object features information can be used for distinguishing scene, foundation
The object of shooting, analyzes the daily object and which place appeared in carries out analysis determination and be currently under which scene, specifically
Including indoor and outdoor, parlor, room, office, meeting room, school, hospital, highway, open air, rock-climbing place etc..The animal is special
Reference ceases available kind, title etc. to animal.The complete of the plot generated may be implemented in the seasonal characteristic information
Property.The weather characteristics information specifically includes fine day, the cloudy day, the rainy day, snows, the description of the associated weathers such as sandstorm.
Usual second feature information will not directly carry in the picture, it is therefore desirable to carry out image analysis in the picture and determine
Second feature information.
Terminal determines the fisrt feature information and second feature information of target image based on image recognition model, may be implemented
The image of shooting is subjected to intelligent measurement identification, including identification personage, scenery, object, animal, season, weather, time, place
Etc., the layout of text is carried out by deep learning, that is, image recognition model, is organized into and is met the image fact and have story logic
Plot.
In order to realize that the logicality and integrality of plot, terminal can be for the target images in target image set
Whether belong to same story to be determined, specific terminal is directed to each target image, judges whether deposit in target image set
In remaining target image of the fisrt feature information association with the target image, when fisrt feature information association can refer to shooting
Between information and shooting location information it is identical, can refer to that shooting location information is identical, shooting time is continuous, if it is shooting time
Continuously, then the time difference between adjacent shooting time is no more than the time difference threshold value of setting, the time difference threshold value can be by
Developer's setting can be and be set by user according to the demand of itself.
If in target image set only including a target image, in target image set necessarily there is no with this
Remaining target image of the fisrt feature information association of one target image；If schemed in target image set including at least two
Picture is then directed to each target image, in target image set there may be, may there is no the fisrt feature with the target image
Remaining target image of information association.
Further, in order to improve the accuracy that plot determines, the whether associated base of fisrt feature information is being determined
On plinth, can also judging second feature information, whether there is also associations, and when simultaneously, fisrt feature information and second feature information are deposited
It determines there is association in association, when judging second feature information with the presence or absence of association, can be according to recognition of face, clothing
Dress ornament and ambient enviroment carry out analysis identification, such as whether the dress for judging that personage wears in the target image can be with it
The clothing of personage decorates matching in remaining target image, that is, judges whether clothes attachment matches；Judge that personage is in the target image
It is no to be matched with personage in remaining target image, that is, judge whether the face recognized matches；Judge around in the target image
Environment, that is, object features information whether can with whether match in the ambient enviroment in remaining target image, i.e. judgement recognizes
Whether object features information matches.Remaining target image, which can be, in this step is different from the mesh in target image set
Remaining target image of logo image can be remaining target image with the fisrt feature information association of the target image.
For each target image, if there is the fisrt feature information association with the target image in target image set
Remaining target image, then a plot can be generated according to the target image and remaining target image, in order to guarantee
The continuity of plot generates plot according to the sequencing of shooting time in the embodiment of the present invention, in order to guarantee therefore
The logicality of thing section, according to the shooting location information and second of the target image and remaining target image in the embodiment of the present invention
Characteristic information generates plot, i.e., according to the shooting location information of the target image and second feature information and remaining mesh
The shooting location information and second feature information of logo image generate plot.
For each target image, if there is no close with the fisrt feature information of the target image in target image set
Remaining target image of connection then directly can generate a plot according to the target image, in order to guarantee story request
Logicality generates plot according to the shooting location information and second feature information of the target image in the embodiment of the present invention.
In order to realize generation plot accuracy, on the basis according to shooting location information and second feature information
On, plot can also be generated according to shooting time information, i.e., when there are remaining associated target image, when according to shooting
Between sequencing, according to shooting location information, shooting time information and the second feature of target image and remaining target image
Information generates a story request；When not depositing remaining associated target image, according to the shooting location information of target image, clap
It takes the photograph temporal information and second feature information generates the request of the first story.
The process for generating story request is completed based on image recognition model, therefore is not illustrated herein.
Plot can be generated according to the target image in target image set in the embodiment of the present invention, realize image
It identifies, and the logicality and continuity of the plot generated.
Embodiment 3:
On the basis of the various embodiments described above, in the embodiment of the present invention, at least one plot packet described in the output
It includes:
For each plot at least one described plot, by the plot and the story feelings are generated
The corresponding output of the target image of section.
Since plot is generated according to the target image in target image set, in order to obtain user more accurately
Know the plot that target image is intended by, therefore can be defeated simultaneously by the target image and plot that generate plot
Out.
Terminal generates plot according to the target image in target image set, and terminal can be according to the story feelings of generation
Section determines the target image for generating the plot, therefore terminal is directed at least one corresponding plot of target image set
In each plot, determine the target image of the generation plot, and then by the plot and generate the event
The corresponding output of the target image of thing section.
Specifically in output, in plot and the identical mark of target image elder generation of the plot can be generated
It is marked, can be in output, plot and the target image for generating the plot are continuously exported, with label
The plot is by the target image generation etc. for being contiguous therewith output, and there are many modes of corresponding output, in the embodiment of the present invention
In be not listed one by one.
By the corresponding output of the target image of plot and generation plot in the embodiment of the present invention, it is able to use family
It can accurately know the plot that target image is intended by.
Embodiment 4:
On the basis of the various embodiments described above, in the embodiment of the present invention, at least one plot packet described in the output
It includes:
At least one described plot is exported in a manner of text segment；And/or
At least one described plot is exported in a manner of voice broadcast.
In order to facilitate the use of different user group, a variety of output sides of plot are provided in the embodiment of the present invention
Formula.
Specific terminal can be and be exported in a manner of text segment and/or with language when exporting at least one plot
The mode of sound casting exports.
For disturbance people or reading, there are the crowds of obstacle, can preferentially select the way of output of voice broadcast.
For in the public situation either poor crowd of understandability ability to express, can preferential text segment it is defeated
Mode out.
Certain user also can choose to be exported in a manner of voice broadcast while being exported in a manner of text segment.
Terminal shows that text and the process of casting voice belong to the prior art, does not repeat them here in embodiments of the present invention.
A variety of way of outputs of plot are provided in the embodiment of the present invention, facilitate the use of different groups, are improved
The popularity and diversity of image-recognizing method user group.
Embodiment 5:
On the basis of the various embodiments described above, in the embodiment of the present invention, if described in being exported in a manner of text segment extremely
A few plot, after at least one plot described in the output, the method also includes:
Judge whether to receive user to the edit requests of text segment, wherein carrying in the edit requests to be edited
Story in the first text set and edited second text set；
If so, the first text set, which is updated to second character set, merges display.
In order to improve the flexibility of image-recognizing method, the embodiment of the present invention also provides editting function for user, for
Family writes text and provides a text template, solves the problems, such as hardly possible of starting writing, user can modify or mend on this basis
It fills and is more suitable for the open of thinking.
When terminal exports corresponding at least one plot of target image set in a manner of text segment, if with
When at least one plot that terminal exports is edited at family, user can be edited the text segment that terminal is shown
Operation, terminal can receive user to the edit requests of text segment at this time, therefore terminal can be by judging whether to receive
User determines whether user edits the text segment of display to the edit requests of text segment.
If the terminal determine that receiving user to the edit requests of text segment, terminal can be identified in edit requests and be carried
Story to be edited in the first text set and edited second text set, which is updated to
Two text set, and the corresponding text segment of updated plot is shown.
If user wants to modify, which can be whole texts or part text in text segment
Word；If user wants to supplement, which can be whole texts in text segment or can see
Work is blank character.
The various embodiments described above are illustrated with a specific embodiment below, as shown in Fig. 2, the terminal has shooting
Function, the terminal shoot photo, and user chooses a photo or multiple pictures in the photo of shooting, and the photo that user chooses can
To regard the target image in the target image set with identification as, terminal is based on image recognition model, turns according to picture recognition
It is melted into explanatory note, that is, determines at least one corresponding plot in target image set, it is automatic to export related verbal description,
At least one plot is exported in the form of text segment, exporting in the form of text segment can edit in order to user
Or supplement.After user compiles function to text segment, the text segment after editing successfully can be used for carrying out correlation analysis or multiple
System is pasted, and is convenient for the user to use, is shared.
The terminal of that embodiment of the invention provides the editting function of user, improves the spirit of image recognition and plot output
Activity.
Embodiment 6:
On the basis of the various embodiments described above, in the embodiment of the present invention, the training process of described image identification model includes:
For each sample image set in training set, the every of the corresponding handmarking of each sample image set is obtained
The corresponding plot of story of the mark and the mark of the story that sample image belongs in a sample image set；
According in the corresponding sample image set of the sample image set, the sample image set got
Sample image ownership story mark and the mark the corresponding plot of story, to described image identification model into
Row training.
Image recognition model provided in an embodiment of the present invention can be the model based on deep learning training completion.
It specifically, include a large amount of sample image set in training set, the sample image collection that training set includes, which is combined into, to be used for
The sample for carrying out model training includes one in each sample image set for each sample image set in training set
Or multiple sample images, terminal can obtain sample in each sample image set of the corresponding handmarking of sample image set
The mark of the story of image ownership, and the corresponding plot of story of mark.
Sample image belongs in the corresponding sample image set of sample image set, sample image set that will acquire
Story mark, and the corresponding plot of story of mark, are trained image recognition model.
The process that model is trained can be realized using the prior art according to sample data, in embodiments of the present invention
It does not repeat them here.
The embodiment of the present invention ensure that in image recognition by being trained to image recognition model, complete according to training
At image recognition model can be realized image is accurately identified.
Embodiment 7:
On the basis of the various embodiments described above, the embodiment of the invention also provides a kind of terminals 300, as shown in figure 3, packet
It includes: processor 301, communication interface 302, memory 303 and communication bus 304, wherein processor 301, communication interface 302 are deposited
Reservoir 303 completes mutual communication by communication bus 304；
It is stored with computer program in the memory 303, when described program is executed by the processor 301, so that
The processor 301 executes following steps:
Obtain target image set to be identified；
The target image set is input in the image recognition model that training is completed in advance, is identified based on described image
Model determines at least one corresponding plot of the target image set；
Export at least one described plot.
Image-recognizing method provided in an embodiment of the present invention is applied to terminal.
The communication bus that above-mentioned terminal is mentioned can be Peripheral Component Interconnect standard (Peripheral Component
Interconnect, PCI) bus or expanding the industrial standard structure (Extended Industry Standard
Architecture, EISA) bus etc..The communication bus can be divided into address bus, data/address bus, control bus etc..For just
It is only indicated with a thick line in expression, figure, it is not intended that an only bus or a type of bus.
Communication interface 302 is for the communication between above-mentioned terminal and other equipment.
Memory may include random access memory (Random Access Memory, RAM), also may include non-easy
The property lost memory (Non-Volatile Memory, NVM), for example, at least a magnetic disk storage.Optionally, memory may be used also
To be storage device that at least one is located remotely from aforementioned processor.
Above-mentioned processor can be general processor, including central processing unit, network processing unit (Network
Processor, NP) etc.；It can also be digital command processor (Digital Signal Processing, DSP), dedicated collection
At circuit, field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hard
Part component etc..
In embodiments of the present invention, it when processor executes the program stored on memory, realizes and is based on having trained in advance
At image recognition model, can determine at least one corresponding plot of target image set to be identified and export, it is real
Show and has helped user's especially special population quick and precisely to obtain the meaning or content wanting to convey in image.
Embodiment 8:
On the basis of the various embodiments described above, the embodiment of the invention also provides a kind of computers to store readable storage medium
Matter is stored with the computer program that can be executed by terminal in the computer readable storage medium, when described program is at the end
When being run on end, so that the terminal realizes following steps when executing:
Obtain target image set to be identified；
The target image set is input in the image recognition model that training is completed in advance, is identified based on described image
Model determines at least one corresponding plot of the target image set；
Export at least one described plot.
Above-mentioned computer readable storage medium can be any usable medium or number that the processor in terminal can access
Such as according to storage equipment, including but not limited to magnetic storage such as floppy disk, hard disk, tape, magneto-optic disk (MO) etc., optical memory
CD, DVD, BD, HVD etc. and semiconductor memory such as ROM, EPROM, EEPROM, nonvolatile memory (NAND
FLASH), solid state hard disk (SSD) etc..
Computer program, computer program are provided in the computer readable storage medium provided in embodiments of the present invention
When being executed by processor, realizes the image recognition model completed based on preparatory training, can determine target image set to be identified
It closes at least one corresponding plot and exports, realize and user's especially special population is helped quick and precisely to obtain image
In want convey the meaning or content.
Fig. 4 is a kind of 400 schematic diagram of pattern recognition device provided in an embodiment of the present invention, is applied to terminal, the device packet
It includes:
Module 401 is obtained, for obtaining target image set to be identified；
Determining module 402, for the target image set to be input in the image recognition model that training is completed in advance,
Based on described image identification model, at least one corresponding plot of the target image set is determined；
Output module 403, for exporting at least one described plot.
The determining module 402 is specifically used for being based on described image identification model, determines mesh in the target image set
The fisrt feature information and second feature information that logo image includes, wherein the fisrt feature information include shooting time information and
Shooting location information；For each target image in the target image set, judge in the target image set whether
In the presence of remaining target image of the fisrt feature information association with the target image；If so, according to the successive suitable of shooting time
Sequence generates an event according to the shooting location information and second feature information in the target image and remaining described target image
Thing section；If not, generating a plot according to the shooting location information and second feature information of the target image.
The second feature information comprises at least one of the following: person characteristic information, scenery characteristic information, object features letter
Breath, animal character information, seasonal characteristic information and weather characteristics information.
The output module 403, specifically for each plot at least one plot for described in, by this
Plot and the corresponding output of the target image for generating the plot.
The output module 403, specifically for exporting at least one described plot in a manner of text segment；With/
Or at least one described plot is exported in a manner of voice broadcast.
The output module 403 is sentenced if being also used to export at least one described plot in a manner of text segment
It is disconnected whether to receive user to the edit requests of text segment, wherein being carried in story to be edited in the edit requests
First text set and edited second text set；If so, the first text set is updated to second text
Word collection merges display.
Described device further include:
Training module 404, for obtaining each sample image set pair for each sample image set in training set
The mark for the story that sample image belongs in each sample image set of the handmarking answered and the story of the mark are corresponding
Plot；According to the corresponding sample graph image set of the sample image set, the sample image set got
The corresponding plot of story of the mark and the mark of the story that sample image belongs in conjunction identifies mould to described image
Type is trained.
The image recognition model completed in the embodiment of the present invention based on preparatory training, can determine target image to be identified
Gather at least one corresponding plot and export, realizes and user's especially special population is helped quick and precisely to obtain figure
The meaning or content conveyed are wanted as in.
For systems/devices embodiment, since it is substantially similar to the method embodiment, so the comparison of description is simple
Single, the relevent part can refer to the partial explaination of embodiments of method.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or an operation are distinguished with another entity or another operation, without necessarily requiring or implying these entities
Or there are any actual relationship or orders between operation.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The application is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although the preferred embodiment of the application has been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications can be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the application range.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (16)
1. a kind of image-recognizing method, which is characterized in that it is applied to terminal, this method comprises:
Obtain target image set to be identified；
The target image set is input in the image recognition model that training is completed in advance, mould is identified based on described image
Type determines at least one corresponding plot of the target image set；
Export at least one described plot.
2. the method as described in claim 1, which is characterized in that it is described to be based on described image identification model, determine the target
At least one corresponding plot of image collection includes:
Based on described image identification model, the fisrt feature information and that target image includes in the target image set are determined
Two characteristic informations, wherein the fisrt feature information includes shooting time information and shooting location information；
For each target image in the target image set, judge to whether there is and the mesh in the target image set
Remaining target image of the fisrt feature information association of logo image；If so, according to the sequencing of shooting time, according to the mesh
Logo image and shooting location information and second feature information in remaining described target image, generate a plot；If
It is no, according to the shooting location information and second feature information of the target image, generate a plot.
3. method according to claim 2, which is characterized in that the second feature information comprises at least one of the following: personage
Characteristic information, scenery characteristic information, object features information, animal character information, seasonal characteristic information and weather characteristics information.
4. method according to claim 2, which is characterized in that at least one plot described in the output includes:
For each plot at least one described plot, by the plot and the plot is generated
The corresponding output of target image.
5. method as described in claim 1 or 4, which is characterized in that at least one plot described in the output includes:
At least one described plot is exported in a manner of text segment；And/or
At least one described plot is exported in a manner of voice broadcast.
6. method as claimed in claim 5, which is characterized in that if exporting at least one described event in a manner of text segment
Thing section, after at least one plot described in the output, the method also includes:
Judge whether to receive user to the edit requests of text segment, wherein carrying event to be edited in the edit requests
The first text set and edited second text set in thing；
If so, the first text set, which is updated to second character set, merges display.
7. method according to claim 1 or 2, which is characterized in that the training process of described image identification model includes:
For each sample image set in training set, each sample of the corresponding handmarking of each sample image set is obtained
The corresponding plot of story of the mark and the mark of the story that sample image belongs in this image collection；
According to sample in the corresponding sample image set of the sample image set, the sample image set got
The corresponding plot of story of the mark and the mark of the story of image ownership, instructs described image identification model
Practice.
8. a kind of pattern recognition device, which is characterized in that be applied to terminal, which includes:
Module is obtained, for obtaining target image set to be identified；
Determining module is based on institute for the target image set to be input in the image recognition model that training is completed in advance
Image recognition model is stated, determines at least one corresponding plot of the target image set；
Output module, for exporting at least one described plot.
9. device as claimed in claim 8, which is characterized in that the determining module is specifically used for identifying based on described image
Model determines target image includes in the target image set fisrt feature information and second feature information, wherein described
Fisrt feature information includes shooting time information and shooting location information；For each target figure in the target image set
Picture judges in the target image set with the presence or absence of remaining target figure with the fisrt feature information association of the target image
Picture；If so, according to the sequencing of shooting time, according to the shooting location in the target image and remaining described target image
Information and second feature information generate a plot；If not, according to the shooting location information of the target image and second
Characteristic information generates a plot.
10. device as claimed in claim 9, which is characterized in that the second feature information comprises at least one of the following: personage
Characteristic information, scenery characteristic information, object features information, animal character information, seasonal characteristic information and weather characteristics information.
11. device as claimed in claim 9, which is characterized in that the output module, be specifically used for for it is described at least one
The plot and the target image for generating the plot are corresponded to output by each plot in plot.
12. the device as described in claim 8 or 11, which is characterized in that the output module, specifically for text segment
Mode exports at least one described plot；And/or at least one described plot is exported in a manner of voice broadcast.
13. device as claimed in claim 12, which is characterized in that the output module, if be also used to text segment
Mode exports at least one described plot, judges whether to receive user to the edit requests of text segment, wherein described
The the first text set and edited second text set in story to be edited are carried in edit requests；If so, will
The first text set is updated to second character set and merges display.
14. device as claimed in claim 8 or 9, which is characterized in that described device further include:
Training module, for obtaining the corresponding people of each sample image set for each sample image set in training set
The corresponding story of story of the mark and the mark of the story of sample image ownership in each sample image set of work label
Plot；According to sample in the corresponding sample image set of the sample image set, the sample image set got
The corresponding plot of story of the mark and the mark of the story of this image ownership, carries out described image identification model
Training.
15. a kind of terminal characterized by comprising processor, communication interface, memory and communication bus, wherein processor,
Communication interface, memory complete mutual communication by communication bus；
It is stored with computer program in the memory, when described program is executed by the processor, so that the processor
Perform claim requires the step of any one of 1~7 the method.
16. a kind of computer readable storage medium, which is characterized in that it is stored with the computer program that can be executed by terminal, when
When described program is run on the terminal, so that the step of terminal perform claim requires any one of 1~7 the method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811196560.3A CN109409423A (en) | 2018-10-15 | 2018-10-15 | A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811196560.3A CN109409423A (en) | 2018-10-15 | 2018-10-15 | A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109409423A true CN109409423A (en) | 2019-03-01 |
Family
ID=65467195
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811196560.3A Pending CN109409423A (en) | 2018-10-15 | 2018-10-15 | A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409423A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902759A (en) * | 2019-03-12 | 2019-06-18 | 上海云从企业发展有限公司 | Pictures describe method and device |
| CN110162639A (en) * | 2019-04-16 | 2019-08-23 | 深圳壹账通智能科技有限公司 | Knowledge figure knows the method, apparatus, equipment and storage medium of meaning |
| CN110378303A (en) * | 2019-07-25 | 2019-10-25 | 杭州睿琪软件有限公司 | Method and system for Object identifying |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095919A (en) * | 2015-09-08 | 2015-11-25 | 北京百度网讯科技有限公司 | Image recognition method and image recognition device |
| CN107274884A (en) * | 2017-02-15 | 2017-10-20 | 赵思聪 | A kind of information acquisition method based on text resolution and phonetic synthesis |
| CN107710188A (en) * | 2015-05-05 | 2018-02-16 | 斯纳普公司 | Automate local story generation and plan exhibition |
| CN107918782A (en) * | 2016-12-29 | 2018-04-17 | 中国科学院计算技术研究所 | A kind of method and system for the natural language for generating description picture material |
| CN108235765A (en) * | 2017-12-05 | 2018-06-29 | 华为技术有限公司 | A kind of display methods and device of story photograph album |
-
2018
- 2018-10-15 CN CN201811196560.3A patent/CN109409423A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107710188A (en) * | 2015-05-05 | 2018-02-16 | 斯纳普公司 | Automate local story generation and plan exhibition |
| CN105095919A (en) * | 2015-09-08 | 2015-11-25 | 北京百度网讯科技有限公司 | Image recognition method and image recognition device |
| CN107918782A (en) * | 2016-12-29 | 2018-04-17 | 中国科学院计算技术研究所 | A kind of method and system for the natural language for generating description picture material |
| CN107274884A (en) * | 2017-02-15 | 2017-10-20 | 赵思聪 | A kind of information acquisition method based on text resolution and phonetic synthesis |
| CN108235765A (en) * | 2017-12-05 | 2018-06-29 | 华为技术有限公司 | A kind of display methods and device of story photograph album |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902759A (en) * | 2019-03-12 | 2019-06-18 | 上海云从企业发展有限公司 | Pictures describe method and device |
| CN110162639A (en) * | 2019-04-16 | 2019-08-23 | 深圳壹账通智能科技有限公司 | Knowledge figure knows the method, apparatus, equipment and storage medium of meaning |
| CN110378303A (en) * | 2019-07-25 | 2019-10-25 | 杭州睿琪软件有限公司 | Method and system for Object identifying |
| CN110378303B (en) * | 2019-07-25 | 2021-07-09 | 杭州睿琪软件有限公司 | Method and system for object recognition |
| US11335087B2 (en) | 2019-07-25 | 2022-05-17 | Hangzhou Glority Software Limited | Method and system for object identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105808782B (en) | A kind of adding method and device of picture tag | |
| CN110119711B (en) | Method and device for acquiring character segments of video data and electronic equipment | |
| US10679063B2 (en) | Recognizing salient video events through learning-based multimodal analysis of visual features and audio-based analytics | |
| CN107993191A (en) | A kind of image processing method and device | |
| JP2019114243A (en) | Imaging device and learning method | |
| CN109074358A (en) | Geographical location related with user interest is provided | |
| CN109409423A (en) | A kind of image-recognizing method, device, terminal and readable storage medium storing program for executing | |
| CN102799890A (en) | Image clustering method | |
| CN110914872A (en) | Navigating video scenes with cognitive insights | |
| JP2017536635A (en) | Picture scene determination method, apparatus and server | |
| CN106528834A (en) | Picture resource pushing method and device based on artificial intelligence | |
| CN110266879A (en) | Broadcast interface display methods, device, terminal and storage medium | |
| CN103870485A (en) | Method and device for achieving augmented reality application | |
| CN112040273B (en) | Video synthesis method and device | |
| CN109040605A (en) | Shoot bootstrap technique, device and mobile terminal and storage medium | |
| WO2023197979A1 (en) | Data processing method and apparatus, and computer device and storage medium | |
| CN114861836B (en) | Model deployment method based on artificial intelligence platform and related equipment | |
| WO2018192245A1 (en) | Automatic scoring method for photo based on aesthetic assessment | |
| CN113627402B (en) | Image identification method and related device | |
| CN106802913A (en) | One kind plays content recommendation method and its device | |
| CN110121033A (en) | Video categorization and device | |
| CN109446946A (en) | A kind of multi-cam real-time detection method based on multithreading | |
| CN116595438A (en) | Picture creation method, device, equipment and storage medium | |
| US11200650B1 (en) | Dynamic image re-timing | |
| WO2021092935A1 (en) | Image data-based message pushing method and device, and computer storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190301 |
|
| RJ01 | Rejection of invention patent application after publication |