CN108416003B - Picture classification method and device, terminal and storage medium - Google Patents
Picture classification method and device, terminal and storage medium Download PDFInfo
- Publication number
- CN108416003B CN108416003B CN201810164391.9A CN201810164391A CN108416003B CN 108416003 B CN108416003 B CN 108416003B CN 201810164391 A CN201810164391 A CN 201810164391A CN 108416003 B CN108416003 B CN 108416003B
- Authority
- CN
- China
- Prior art keywords
- target user
- classification
- label
- picture
- current
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/51—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a picture classification method and device, a terminal and a storage medium, wherein the method comprises the steps of obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user; predicting a current classification label set of the target user according to current characteristic information of the target user by using a label screening model; and identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set. The embodiment of the invention can realize the effects of enriching the existing picture classification method and meeting the personalized classification requirements of users.
Description
Technical Field
The embodiment of the invention relates to the technical field of computers, in particular to a picture classification method and device, a terminal and a storage medium.
Background
At present, the number of pictures stored in various terminals is increasing, and it is very important to effectively manage the pictures. For example, in the existing mobile terminal album classification process, the classification management of albums is mostly realized by using a default label classification scheme of the mobile terminal, sometimes, a user needs to manually define labels, photos organized according to different dimensions are manually managed, and one photo can only appear in one type of label classification. Based on the above classification scheme, the following drawbacks mainly exist:
1) the dimensions of the album label are limited: generally, the photo album classification method only comprises simple labels such as human faces or positioning, and the simple labels are very limited in the face of diversified use scenes of different users.
2) The scheme lacks extensibility: the classification scheme is single, and the classification expansion cannot be performed on the current classification scheme, for example: the time dimension can not be expanded according to the special memorial days such as legal holidays, wedding memorial days or custom time.
3) The scheme lacks intelligence: the scheme can not carry out personalized classification according to the characteristics of the user, can not reach the classification requirements of thousands of people and thousands of faces, and can not accurately understand the classification requirements of the user.
Therefore, how to enrich the existing image classification scheme and realize the personalized classification of the images according to different user requirements still remains a problem to be solved in the image management process.
Disclosure of Invention
The embodiment of the invention provides a picture classification method and device, a terminal and a storage medium, which are used for enriching the existing picture classification method and meeting the effect of individual classification requirements of users.
In a first aspect, an embodiment of the present invention provides a method for classifying pictures, where the method includes:
the method comprises the steps of obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user;
predicting a current classification label set of the target user according to current characteristic information of the target user by using the label screening model;
and identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
In a second aspect, an embodiment of the present invention further provides an image classification apparatus, where the apparatus includes:
the label screening model obtaining module is used for obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user;
the current classification label set prediction module is used for predicting the current classification label set of the target user according to the current characteristic information of the target user by utilizing the label screening model;
and the picture classification module is used for identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
In a third aspect, an embodiment of the present invention further provides a terminal, including:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a method of picture classification as in any embodiment of the invention.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the picture classification method according to any embodiment of the present invention.
According to the method and the device, a label screening model obtained by pre-training according to historical characteristic information of a target user is obtained, a current classification label set of the target user is predicted according to the current characteristic information of the target user, and then the picture of the target user is identified, so that at least one picture classification set corresponding to each classification label in the current classification label set is obtained. The embodiment of the invention solves the problems that the image classification method in the prior art is single and can not realize personalized classification for users, and realizes the effects of enriching the prior image classification method and meeting the personalized classification requirements of the users.
Drawings
Fig. 1 is a flowchart of a picture classification method according to an embodiment of the present invention;
fig. 2 is a flowchart of a picture classification method according to a second embodiment of the present invention;
fig. 3 is a flowchart of a picture classification method according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of an image classification apparatus according to a fourth embodiment of the present invention;
fig. 5 is a schematic structural diagram of a terminal according to a fifth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a picture classification method according to an embodiment of the present invention, where the embodiment is applicable to a case of classifying pictures, and the method may be executed by a picture classification device, and the device may be implemented in a software and/or hardware manner, and may be integrated in a terminal, for example, an intelligent product such as a computer and a mobile terminal. As shown in fig. 1, the method specifically includes:
s110, a label screening model of the target user is obtained, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user.
The target user can leave a corresponding operation trace in the process of using the terminal, the terminal can store data corresponding to the operation trace, and then the characteristic information of the target user can be obtained based on data analysis. The characteristic information comprises interest and preference information of the target user and the like, namely the operation trace reflects the interest and preference of the target user to a certain extent. For example, the target user may analyze that the target user may be interested in the pet cat by searching and browsing pictures and videos about the pet cat on the internet for a relatively large number of times during the use of the terminal. The historical characteristic information can be obtained by analyzing the operation trace data of the target user accumulated at regular intervals, and the specific regular time can be set according to the classification requirement of the target user. Based on the obtained historical characteristic information, the terminal can train by using a machine learning method to obtain a label screening model, so that the classification label of the tendency of the target user in image classification is predicted, the phenomenon that the user manually sets the classification label is avoided, and the intellectualization of image classification is improved.
And S120, predicting the current classification label set of the target user according to the current characteristic information of the target user by using the acquired label screening model.
The current characteristic information of the target user is a direct reaction to the current state of the target user, and based on the current characteristic information, the current classification label set of the target user can be predicted by using a label screening model. The current classification label set comprises a time label, a place label, a human face label, a plant label, an animal label, a weather label, an action label, a color label, a combination of at least two of the above label types and the like, wherein each type of label can be specifically subdivided, for example, the time label comprises a legal holiday label, a birthday label, a commemorative day label, a user-defined marking time label and the like, and the place label comprises place labels of different countries and/or different cities of the same country. Illustratively, the current feature information of the target user reflects that the target user is interested in oriental cherry, and the current classification label set predicted by the label screening model has a combined label like oriental cherry.
The current classification label set of the target user is predicted to automatically generate the picture classification labels, so that the target user can be helped to efficiently manage pictures in the terminal, personalized picture classification aiming at the characteristics of the user is realized, in addition, the current classification label set can comprise various combined labels, the relevance among different labels is increased by the combined labels, the combined labels can correspond to different use scenes, the multi-dimensional extension of the limited picture classification labels in the prior art is realized, and the problems that in the prior art, single label classification schemes are isolated from each other, and the labels cannot be flexibly combined to realize classification are solved.
S130, identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
After the current classification label set of the target user is determined, the picture of the target user can be identified, and the picture is classified under the corresponding label through matching of the picture characteristic information and the label information in the label set, so that the picture classification set is obtained. For example, corresponding to a combined tag of a place tag and a plant tag, namely, tokyo cherry blossom, the terminal can identify all pictures of a target user, and classify the pictures which simultaneously meet two feature information of tokyo cherry blossom into a tokyo cherry blossom picture set. The target user pictures comprise pictures stored in a terminal local photo album and also comprise pictures in a terminal cloud photo album. When the pictures in the terminal cloud photo album are classified, the pictures in the cloud photo album need to be classified by means of network communication transmission.
After the current classification tag set of the target user is predicted, the terminal can establish a certain cache for each tag for temporarily storing the pictures corresponding to the tag, and after all the pictures are classified, data in the caches storing the classified pictures are respectively transferred to a local storage space of the terminal for storage, and the caches without the classified pictures are released; the terminal may also perform image identification and establish a corresponding image classification set, which is not limited in this embodiment. And (4) directly storing the classified picture set in the cloud photo album by picture classification. By automatically labeling and classifying the pictures, great convenience is brought to searching the pictures for the target user, and the time efficiency of searching and searching the pictures is improved.
On the basis of the technical scheme, further, the training process of the label screening model comprises the following steps:
acquiring operation behavior data of a target user on a terminal; optionally, the operation behavior data includes operation behavior data of the target user on pictures, characters and/or audios and videos on the terminal;
analyzing and obtaining behavior parameters of the target user according to the obtained operation behavior data, wherein the behavior parameters are used for representing the characteristics of the target user;
combining the acquired behavior parameters with the user image of the target user to serve as the characteristics of the target user;
and taking the target user characteristics as input, taking the classification label labeling result of the target user characteristics as output, and training by using a machine learning method to obtain a label screening model.
The operation behavior data of the target user may be operation behavior data of pictures, texts and/or audios and videos, specifically, for example, operation behavior data of pictures and/or videos recently taken by the target user, pictures, texts and/or audios and videos browsed on the internet, or pictures shared or complied with on a public platform, and these data are equivalent to a statistical representation of interests and preferences of the target user. The behavior parameters representing the characteristics of the target user are obtained by performing image recognition, keyword extraction, semantic analysis or audio/video analysis on the operation behavior data, for example, the behavior parameters may include parameters of behavior representatives (such as sharing, praise or browsing), data type parameters (such as shared pictures, texts or voices), semantic parameters of shared data (such as cats), and the like.
The user portrait is a virtual representation formed by the network product supplier according to different user differences, such as a hundred-degree user portrait, behavior parameters corresponding to operation behavior data of a target user and the user portrait are combined to serve as training data of the label screening model, and accuracy of model training can be improved.
Illustratively, a target user frequently searches or browses pictures and articles related to cats recently, the terminal takes behavior parameters representing the interest of the target user in the cats and the user portrait as input, a tag screening model is obtained through training, the output result of the tag screening model may be a classification tag set including a tag "lovely pet" or a tag "cat", and when pictures related to the cats or pets are stored in a photo album of the mobile terminal of the user, the pictures and the articles related to the cats are classified into the set including the tag "lovely pet" or the tag "cat".
According to the technical scheme, the method comprises the steps of obtaining a label screening model obtained by pre-training according to historical characteristic information of a target user, predicting a current classification label set of the target user according to the current characteristic information of the target user, identifying a picture of the target user, and obtaining at least one picture classification set corresponding to each classification label in the current classification label set, so that the problems that in the prior art, a picture classification method is single and personalized classification cannot be realized for the user are solved, the pictures of the target user are efficiently and accurately classified, the existing picture classification methods are enriched, the personalized classification requirements of the user are met, and the picture classification result is made to be fit with the characteristics and interests of the target user; in addition, the current classification label set can comprise various combined labels, the combined labels increase the relevance among different labels, can correspond to different use scenes, solves the problems that single label classification schemes are isolated from each other and labels cannot be flexibly combined to realize classification in the prior art, and realizes multi-dimensional extension of limited picture classification labels in the prior art.
Example two
Fig. 2 is a flowchart of a picture classification method according to a second embodiment of the present invention, which is further optimized based on the above-mentioned embodiments. As shown in fig. 2, the method specifically includes:
s210, a label screening model of the target user is obtained, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user.
And S220, predicting the current classification label set of the target user according to the current characteristic information of the target user by using the acquired label screening model.
And S230, determining at least one image recognition model corresponding to the current classification label set according to the types of the classification labels in the current classification label set.
Because the current classification label set comprises various labels, and the pictures corresponding to different types of labels can be identified through different identification algorithms, different types of labels can correspond to different image identification models, for example, the picture corresponding to a face type label can be identified through the face identification algorithm, the picture corresponding to a place type label can be identified through the place matching algorithm, and the picture corresponding to the face type label and the picture corresponding to the place type label correspond to different image identification models.
Optionally, the training process of the image recognition model includes: and taking the historical picture set with the classification label labels as input, taking the labeled classification labels as output, and training by using a machine learning method to obtain an image recognition model based on different classification labels.
S240, identifying the picture of the target user by using at least one image identification model to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
The images of the target user are identified by using different image identification models, so that the accuracy of image identification can be ensured, and the accuracy of image classification can be further ensured.
Optionally, the method further includes updating the label screening model and the image recognition model periodically through training.
The terminal can ensure the dynamic property of the image classification method in the embodiment by regularly updating the label screening model and the image identification model. In addition, in the picture classification process, the used models are the latest models no matter the current classification label set of the target user is predicted or the picture of the target user is identified by using the image identification model, so that the use experience of the user is ensured.
The technical scheme of the embodiment predicts the current classification label set of the target user according to the current characteristic information of the target user by acquiring the label screening model obtained by pre-training according to the historical characteristic information of the target user, then, the image recognition models corresponding to different types of labels are used for recognizing and classifying the pictures of the target user, the problems that the picture classification method in the prior art is single and cannot realize personalized classification for the user are solved, the pictures of the target user are efficiently and accurately classified, and enriches the existing picture classification methods and meets the effect of the personalized classification requirements of users, in addition, in the picture classification process, the problem that the picture classification method in the prior art cannot perform dynamic modification and optimization is solved through the regular updating and perfection of the label screening model and the image recognition model, and the use experience of a user is guaranteed.
EXAMPLE III
Fig. 3 is a flowchart of a picture classification method provided in the third embodiment of the present invention, and the third embodiment is further optimized based on the foregoing embodiments. As shown in fig. 3, the method specifically includes:
s310, a tag screening model of the target user is obtained from the cloud, wherein the tag screening model is obtained based on cloud data training of the target user and is updated in the cloud.
Based on the cloud data of the target user, the training of the label screening model is realized at the cloud, the program running pressure during the model training of the terminal can be relieved, and the phenomena of system blockage and the like caused by more task processes of the terminal are avoided. The cloud data is information data which is collected by the server from the terminal and can represent characteristics of the target user, and comprises behavior parameters and a user portrait of the target user. And the cloud end can store the label screening model after finishing the training of the label screening model, and regularly updates the label screening model according to the change of the cloud end data of the target user. And after receiving an acquisition request of the terminal for the label screening model, issuing the current label screening model to the terminal. Meanwhile, the cloud end can judge whether the current tag screening model is updated or not so as to ensure that the latest tag screening model is issued to the terminal.
And S320, predicting the current classification label set of the target user according to the current characteristic information of the target user by using the acquired label screening model.
The terminal acquires a label screening model which is sent by the cloud and is trained, stores the model locally, and can predict the current classification label set of the target user according to the current characteristic information of the target user, wherein the prediction can be online prediction or offline prediction and is not limited by the current network state of the terminal.
S330, determining at least one image recognition model corresponding to the current classification tag set from the cloud according to the types of the classification tags in the current classification tag set, wherein the image recognition model is obtained through training in the cloud and is updated in the cloud.
Similar to the label screening model, the training, the saving and the updating of the image recognition model are all carried out at the cloud end, and the occupation of the memory space of the terminal can be avoided. And after the predicted current classification label set is determined, the terminal sends a confirmation request of the image recognition model to the cloud, the cloud issues the image recognition model corresponding to the specific type label to the terminal according to the confirmation request, and the terminal acquires the image recognition model and stores the image recognition model locally. Similarly, before issuing the model, the cloud judges whether the current image recognition model is updated or not so as to ensure that the latest image recognition model is issued to the terminal.
S340, identifying the picture of the target user by using at least one image identification model to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
It should be noted that when the pictures of the terminal are stored in the cloud album, the pictures in the cloud album of the target user can be identified and classified by directly using network communication and using the tag screening model and the image identification model of the cloud.
According to the technical scheme, the tag screening model is obtained from the cloud, the current classification tag set of the target user is predicted according to the current characteristic information of the target user, and then the image recognition model obtained from the cloud is used for recognizing and classifying the image of the target user, so that the problems that in the prior art, the image classification method is single, and personalized classification cannot be realized for the user are solved, the effects of enriching the existing image classification method and meeting the personalized classification requirements of the user are realized, the task pressure of terminal system operation is relieved by completing training and updating of the tag screening model and the image recognition model at the cloud, the smoothness of terminal system operation is guaranteed, the model is issued from the cloud to the terminal, the effect of offline classification of the image can be realized, and the constraint of the terminal network connection state on image classification is eliminated.
Example four
Fig. 4 is a schematic structural diagram of an image classification device according to a fourth embodiment of the present invention, which is applicable to classifying images. The image classification device provided by the embodiment of the invention can execute the image classification method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method. As shown in fig. 4, the apparatus includes a tag screening model obtaining module 410, a current classification tag set predicting module 420 and a picture classifying module 430, wherein:
a tag screening model obtaining module 410, configured to obtain a tag screening model of a target user, where the tag screening model is obtained by pre-training according to historical feature information of the target user and is used to predict a classification tag of the target user;
a current classification tag set prediction module 420, configured to predict, according to the current feature information of the target user, a current classification tag set of the target user by using the obtained tag screening model;
the image classification module 430 is configured to identify an image of a target user, and obtain at least one image classification set corresponding to each classification tag in a current classification tag set.
Further, the picture classification module 430 includes:
the image identification model set determining unit is used for determining at least one image identification model corresponding to the current classification label set according to the types of the classification labels in the current classification label set;
and the picture classification unit is used for identifying the picture of the target user by using at least one image identification model to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
Optionally, the apparatus further comprises:
the label screening model training module is used for training to obtain the label screening model; wherein, this label screening model training module includes:
the operation behavior data acquisition unit is used for acquiring operation behavior data of a target user on the terminal;
the behavior parameter acquiring unit is used for analyzing and obtaining behavior parameters of the target user according to the acquired operation behavior data, wherein the behavior parameters are used for representing the characteristics of the target user;
the target user characteristic acquisition unit is used for combining the acquired behavior parameters with the user image of the target user to serve as target user characteristics;
and the model training unit is used for taking the target user characteristics as input, taking the classification label labeling result of the target user characteristics as output, and training by utilizing a machine learning method to obtain a label screening model.
Optionally, the operation behavior data acquiring unit is specifically configured to: and acquiring the operation behavior data of the target user on the pictures, the characters and/or the audio and video on the terminal.
Optionally, the apparatus further comprises:
and the updating module is used for updating the label screening model and the image recognition model regularly through training.
According to the method and the device, a label screening model obtained by pre-training according to historical characteristic information of a target user is obtained, a current classification label set of the target user is predicted according to the current characteristic information of the target user, and then the picture of the target user is identified, so that at least one picture classification set corresponding to each classification label in the current classification label set is obtained. The embodiment of the invention solves the problems that the image classification method in the prior art is single and can not realize personalized classification for users, and realizes the effects of enriching the prior image classification method and meeting the personalized classification requirements of the users.
EXAMPLE five
Fig. 5 is a schematic structural diagram of a terminal according to a fifth embodiment of the present invention. Fig. 5 illustrates a block diagram of an exemplary terminal 512 suitable for use in implementing embodiments of the present invention. The terminal 512 shown in fig. 5 is only an example and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.
As shown in fig. 5, the terminal 512 is represented in the form of a general-purpose terminal. The components of the terminal 512 may include, but are not limited to: one or more processors 516, a storage device 528, and a bus 518 that couples the various system components including the storage device 528 and the processors 516.
The terminal 512 typically includes a variety of computer system readable media. Such media can be any available media that is accessible by terminal 512 and includes both volatile and nonvolatile media, removable and non-removable media.
A program/utility 540 having a set (at least one) of program modules 542 may be stored, for example, in storage 528, such program modules 542 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may include an implementation of a network environment. The program modules 542 generally perform the functions and/or methods of the described embodiments of the invention.
The terminal 512 may also communicate with one or more external devices 514 (e.g., keyboard, pointing device, display 524, etc.), with one or more devices that enable a user to interact with the terminal 512, and/or with any devices (e.g., network card, modem, etc.) that enable the terminal 512 to communicate with one or more other computing devices. Such communication may occur via input/output (I/O) interfaces 522. Also, the terminal 512 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public Network such as the internet) via the Network adapter 520. As shown in fig. 5, the network adapter 520 communicates with the other modules of the terminal 512 via the bus 518. It should be appreciated that although not shown, other hardware and/or software modules may be used in conjunction with the terminal 512, including but not limited to: microcode, device drivers, Redundant processors, external disk drive Arrays, RAID (Redundant Arrays of Independent Disks) systems, tape drives, and data backup storage systems, among others.
The processor 516 executes various functional applications and data processing by running a program stored in the storage device 528, for example, implementing a picture classification method provided by an embodiment of the present invention, the method includes:
the method comprises the steps of obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user;
predicting a current classification label set of the target user according to current characteristic information of the target user by using the label screening model;
and identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
EXAMPLE six
The sixth embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for classifying pictures, where the method includes:
the method comprises the steps of obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user;
predicting a current classification label set of the target user according to current characteristic information of the target user by using the label screening model;
and identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a Read-Only Memory (ROM), an Erasable Programmable Read-Only Memory (EPROM, or flash Memory), an optical fiber, a portable compact disc Read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, Radio Frequency (RF), etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or terminal. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (12)
1. A picture classification method is characterized by comprising the following steps:
the method comprises the steps of obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user; the historical characteristic information comprises: behavior parameters of a target user and a user portrait, wherein the behavior parameters comprise parameters of behavior representatives, data type parameters and data semantic parameters;
predicting a current classification label set of the target user according to current characteristic information of the target user by using the label screening model; wherein the current classification tag set comprises a single tag and a combined tag;
and identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
2. The method of claim 1, wherein the training process of the tag screening model comprises:
acquiring operation behavior data of a target user on a terminal;
analyzing and obtaining behavior parameters of the target user according to the operation behavior data, wherein the behavior parameters are used for representing the characteristics of the target user;
combining the behavior parameters with the user image of the target user to serve as the characteristics of the target user;
and taking the target user characteristics as input, taking the classification label labeling result of the target user characteristics as output, and training by using a machine learning method to obtain the label screening model.
3. The method according to claim 2, wherein the operation behavior data comprises operation behavior data of a target user on pictures, texts and/or audios and videos on the terminal.
4. The method of claim 1, wherein the identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification tag in the current classification tag set comprises:
determining at least one image recognition model corresponding to the current classification label set according to the types of the classification labels in the current classification label set;
and identifying the picture of the target user by using the at least one image identification model to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
5. The method of claim 4, further comprising:
and updating the label screening model and the image recognition model through training periodically.
6. An apparatus for classifying pictures, comprising:
the label screening model obtaining module is used for obtaining a label screening model of a target user, wherein the label screening model is obtained by pre-training according to historical characteristic information of the target user and is used for predicting a classification label of the target user; the historical characteristic information comprises: behavior parameters of a target user and a user portrait, wherein the behavior parameters comprise parameters of behavior representatives, data type parameters and data semantic parameters;
the current classification label set prediction module is used for predicting the current classification label set of the target user according to the current characteristic information of the target user by utilizing the label screening model; wherein the current classification tag set comprises a single tag and a combined tag;
and the picture classification module is used for identifying the picture of the target user to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
7. The apparatus of claim 6, further comprising:
the label screening model training module is used for training to obtain the label screening model; wherein, the label screening model training module comprises:
the operation behavior data acquisition unit is used for acquiring operation behavior data of a target user on the terminal;
the behavior parameter acquiring unit is used for analyzing and obtaining behavior parameters of the target user according to the operation behavior data, wherein the behavior parameters are used for representing the characteristics of the target user;
the target user characteristic acquisition unit is used for combining the behavior parameters with the user image of the target user to serve as target user characteristics;
and the model training unit is used for taking the target user characteristics as input, taking the classification label labeling result of the target user characteristics as output, and training by utilizing a machine learning method to obtain the label screening model.
8. The apparatus according to claim 7, wherein the operation behavior data obtaining unit is specifically configured to: and acquiring the operation behavior data of the target user on the pictures, the characters and/or the audio and video on the terminal.
9. The apparatus of claim 6, wherein the picture classification module comprises:
the image identification model set determining unit is used for determining at least one image identification model corresponding to the current classification label set according to the type of the classification label in the current classification label set;
and the picture classification unit is used for identifying the picture of the target user by using the at least one image identification model to obtain at least one picture classification set corresponding to each classification label in the current classification label set.
10. The apparatus of claim 9, further comprising:
and the updating module is used for updating the label screening model and the image recognition model regularly through training.
11. A terminal, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a picture classification method as claimed in any one of claims 1 to 5.
12. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method for picture classification according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810164391.9A CN108416003B (en) | 2018-02-27 | 2018-02-27 | Picture classification method and device, terminal and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810164391.9A CN108416003B (en) | 2018-02-27 | 2018-02-27 | Picture classification method and device, terminal and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108416003A CN108416003A (en) | 2018-08-17 |
| CN108416003B true CN108416003B (en) | 2021-12-21 |
Family
ID=63129115
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810164391.9A Active CN108416003B (en) | 2018-02-27 | 2018-02-27 | Picture classification method and device, terminal and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108416003B (en) |
Families Citing this family (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110955790A (en) * | 2018-09-26 | 2020-04-03 | Oppo广东移动通信有限公司 | Picture classification method and device, computer-readable storage medium and electronic equipment |
| CN109446783B (en) * | 2018-11-16 | 2023-07-25 | 山东浪潮科学研究院有限公司 | Image recognition efficient sample collection method and system based on machine crowdsourcing |
| CN109597912A (en) * | 2018-12-05 | 2019-04-09 | 上海碳蓝网络科技有限公司 | Method for handling picture |
| CN111027557B (en) * | 2019-03-11 | 2024-03-19 | 广东小天才科技有限公司 | Subject identification method based on subject image and electronic equipment |
| CN110110750B (en) * | 2019-03-29 | 2021-03-05 | 广州思德医疗科技有限公司 | Original picture classification method and device |
| CN110390033B (en) * | 2019-07-25 | 2023-04-21 | 腾讯科技（深圳）有限公司 | Training method and device for image classification model, electronic equipment and storage medium |
| CN112580681B (en) * | 2019-09-30 | 2022-02-01 | 北京星选科技有限公司 | User classification method and device, electronic equipment and readable storage medium |
| CN110874420A (en) * | 2019-11-09 | 2020-03-10 | 上海光数信息科技有限公司 | Picture information processing system and method |
| CN111143568A (en) * | 2019-12-31 | 2020-05-12 | 郑州工程技术学院 | Method, device and equipment for buffering during paper classification and storage medium |
| CN111274427A (en) * | 2020-01-08 | 2020-06-12 | 上海擎感智能科技有限公司 | Picture processing method and device and computer storage medium |
| CN111399778A (en) * | 2020-03-17 | 2020-07-10 | 西安石油大学 | Picture file cleaning method and device |
| CN113496236B (en) * | 2020-03-20 | 2024-05-24 | 北京沃东天骏信息技术有限公司 | User tag information determining method, device, equipment and storage medium |
| CN112765386A (en) * | 2020-06-14 | 2021-05-07 | 黄雨勤 | Information management method and system based on big data and Internet and cloud server |
| CN113936206A (en) * | 2020-07-08 | 2022-01-14 | 华为技术有限公司 | Method, medium, and system for classifying image on terminal device |
| CN111966850A (en) * | 2020-07-21 | 2020-11-20 | 珠海格力电器股份有限公司 | Picture screening method and device, electronic equipment and storage medium |
| CN112288793B (en) * | 2020-11-06 | 2023-05-26 | 洛阳语音云创新研究院 | Method and device for detecting backfat of livestock individuals, electronic equipment and storage medium |
| CN114047996A (en) * | 2021-11-25 | 2022-02-15 | 维沃移动通信有限公司 | Wallpaper setting method, wallpaper setting device and electronic equipment |
| CN115019218B (en) * | 2022-08-08 | 2022-11-15 | 阿里巴巴（中国）有限公司 | Image processing method and processor |
| CN116127108A (en) * | 2023-02-20 | 2023-05-16 | 广东德融汇科技有限公司 | Album generation method and device based on image recognition and storage medium |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8478052B1 (en) * | 2009-07-17 | 2013-07-02 | Google Inc. | Image classification |
| CN103995889B (en) * | 2014-06-03 | 2017-11-03 | 广东欧珀移动通信有限公司 | Picture classification method and device |
| CN105760448B (en) * | 2016-02-03 | 2019-11-15 | 北京金山安全软件有限公司 | Picture processing method and device and electronic equipment |
| CN106055617A (en) * | 2016-05-26 | 2016-10-26 | 乐视控股（北京）有限公司 | Data pushing method and device |
| CN106445995A (en) * | 2016-07-18 | 2017-02-22 | 腾讯科技（深圳）有限公司 | Picture classification method and apparatus |
| CN106909629B (en) * | 2017-01-26 | 2020-05-19 | 武汉奇米网络科技有限公司 | Method and system for individually recommending pit bit labels |
| CN107341245B (en) * | 2017-07-06 | 2020-08-21 | 阿里巴巴（中国）有限公司 | Data processing method and device and server |
-
2018
- 2018-02-27 CN CN201810164391.9A patent/CN108416003B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108416003A (en) | 2018-08-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108416003B (en) | Picture classification method and device, terminal and storage medium | |
| CN108319723B (en) | Picture sharing method and device, terminal and storage medium | |
| US20200183977A1 (en) | Providing relevant cover frame in response to a video search query | |
| CN108052577B (en) | Universal text content mining method, device, server and storage medium | |
| US10970334B2 (en) | Navigating video scenes using cognitive insights | |
| CN112203122A (en) | Artificial intelligence-based similar video processing method and device and electronic equipment | |
| US10353943B2 (en) | Computerized system and method for automatically associating metadata with media objects | |
| CN110019616B (en) | POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof | |
| EP3109775A1 (en) | Multimedia content providing method and device | |
| CN106611015B (en) | Label processing method and device | |
| CN110598157A (en) | Target information identification method, device, equipment and storage medium | |
| CN109947989B (en) | Method and apparatus for processing video | |
| CN110633423B (en) | Target account identification method, device, equipment and storage medium | |
| CN112650867B (en) | Picture matching method and device, electronic equipment and storage medium | |
| WO2024051609A1 (en) | Advertisement creative data selection method and apparatus, model training method and apparatus, and device and storage medium | |
| CN113033682B (en) | Video classification method, device, readable medium and electronic equipment | |
| CN112925905B (en) | Method, device, electronic equipment and storage medium for extracting video subtitles | |
| CN104520848A (en) | Searching for events by attendants | |
| CN112732949A (en) | Service data labeling method and device, computer equipment and storage medium | |
| CN112188311A (en) | Method and apparatus for determining video material of news | |
| CN111382281A (en) | Recommendation method, device, equipment and storage medium based on content of media object | |
| US11437038B2 (en) | Recognition and restructuring of previously presented materials | |
| WO2023124793A1 (en) | Image pushing method and device | |
| CN111127057B (en) | Multi-dimensional user portrait recovery method | |
| CN110704650A (en) | OTA picture tag identification method, electronic device and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |