CN109829073A - A kind of method and device of picture search - Google Patents
A kind of method and device of picture search Download PDFInfo
- Publication number
- CN109829073A CN109829073A CN201811640366.XA CN201811640366A CN109829073A CN 109829073 A CN109829073 A CN 109829073A CN 201811640366 A CN201811640366 A CN 201811640366A CN 109829073 A CN109829073 A CN 109829073A
- Authority
- CN
- China
- Prior art keywords
- identifier
- image
- data
- preset model
- mark
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- 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
- 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
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Library & Information Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Processing Or Creating Images (AREA)
Abstract
The present invention provides a kind of method and devices of picture search.The described method includes: the inquiry request of user is received, wherein including target image and screening conditions in the inquiry request；It obtains the characteristic value of the target image, and the characteristic value of the target image is input to the first identifier for obtaining the data of N before similarity ranking in preset model；Determine second identifier corresponding with the first identifier；The second identifier is input in destination server to obtain first structure data；The first structure data are screened to obtain the second structural data according to the screening conditions；Wherein include that third identifies in second structural data, result images is obtained according to third mark and the preset model, and return to the result images to the user.The technical solution provided through the invention can save memory space while promoting search efficiency.
Description
Technical field
The present invention relates to internet area, in particular to a kind of method and device of picture search.
Background technique
With the development of science and technology, human information has been in explosive growth, and many search engine manufacturers have to add
A large amount of server carries out data storage.
Correspondingly, hitting target the time of file since data volume is increasing, when user being caused to scan for increasingly
Long, search efficiency is low.
Summary of the invention
The embodiment of the invention provides a kind of method and device of picture search, by using method provided by the invention,
The structural data of image and unstructured data are stored separately, when carrying out picture search, pass through unstructured number
According to similarity calculation and structural data screening can quick definitive result image, while promoting search efficiency,
Save memory space.
First aspect present invention discloses a kind of method of picture search, which comprises
The inquiry request of user is received, wherein including target image and screening conditions in the inquiry request；
The characteristic value of the target image is obtained, and the characteristic value of the target image is input in preset model to obtain
Take the first identifier of the data of N before similarity ranking；Wherein, N is positive integer；
Determine second identifier corresponding with the first identifier；
The second identifier is input in destination server to obtain first structure data；
The first structure data are screened to obtain the second structural data according to the screening conditions；Wherein
It include that third identifies in second structural data, the third is identified as some or all of second identifier；
Result images are obtained according to third mark and the preset model, and return to the result figure to the user
Picture.
It is optionally, described that result images are obtained according to third mark and the preset model, comprising:
Determine it is corresponding with third mark the 4th identify, wherein the described 4th be identified as the part of first identifier or
All；
4th mark is input in the preset model to obtain result images.
Optionally, before the inquiry request for receiving user, the method also includes:
When receiving the request of batch increase image, the structural data and unstructured number of the image of batch are obtained
According to；
The structural data of the image of the batch is stored into the destination server；And
The unstructured data of the image of the batch is stored into the preset model.
Optionally, the unstructured data of the image data by the batch is stored into the preset model, packet
It includes:
File destination is generated according to the unstructured data of the image of the batch；
The file destination is loaded using the preset model.
Optionally, determination second identifier corresponding with the first identifier, comprising:
The first identifier and the mapping table prestored are matched to obtain corresponding with the first identifier described the
Two marks；Wherein, the mapping table table is stored with the mapping relations of the first identifier and second identifier, and the first identifier is
The mark of the image stored in preset model, the second identifier are sequence mark of the described image in the destination server
Know.
Second aspect of the present invention discloses a kind of device of picture search, and described device includes:
Receiving unit, for receiving the inquiry request of user, wherein including target image and screening in the inquiry request
Condition；
Acquiring unit, for obtaining the characteristic value of the target image；
Input unit, before the characteristic value of the target image is input in preset model to obtain similarity ranking
The first identifier of the data of N；Wherein, N is positive integer；
Determination unit, for determining second identifier corresponding with the first identifier；
The input unit, for the second identifier being input in destination server to obtain first structure number
According to；
Screening unit, for being screened the first structure data to obtain the second knot according to the screening conditions
Structure data；Wherein in second structural data include third identify, the third be identified as second identifier part or
All；
Acquiring unit, for obtaining result images according to third mark and the preset model；
Return unit, for returning to the result images to the user.
Optionally, the acquiring unit is specifically used for determining the 4th mark corresponding with third mark, wherein institute
It states the 4th and is identified as some or all of first identifier；4th mark is input in the preset model to obtain result
Image.
Optionally, described device further includes the first storage unit and the second storage unit；
The acquiring unit is also used to obtain the structure of the image of batch when receiving the request of batch increase image
Change data and unstructured data；
First storage unit, for storing the structural data of the image of the batch to the destination server
In；
Second storage unit, for storing the unstructured data of the image of the batch to the preset model
In.
Optionally, second storage unit, the unstructured data for the image according to the batch generate target
File；The file destination is loaded using the preset model.
Optionally, the determination unit, specifically for matching the first identifier with the mapping table prestored to obtain
Take the second identifier corresponding with the first identifier；Wherein, the mapping table table is stored with the first identifier and second
The mapping relations of mark, the first identifier are the mark of the image stored in preset model, and the second identifier is the figure
As the sequence identifier in the destination server.
As can be seen that the inquiry request of user is received, wherein in the inquiry request in the scheme of the embodiment of the present invention
Including target image and screening conditions；The characteristic value of the target image is obtained, and the characteristic value of the target image is inputted
The first identifier of the data of N before similarity ranking is obtained into preset model；Wherein, N is positive integer；It determines and described first
Identify corresponding second identifier；The second identifier is input in destination server to obtain first structure data；According to
The screening conditions screen to obtain the second structural data the first structure data；Wherein second structure
Changing in data includes that third identifies, and the third is identified as some or all of second identifier；According to third mark and institute
It states preset model and obtains result images, and return to the result images to the user.The technical solution provided through the invention,
By using method provided by the invention, the structural data of image and unstructured data are stored separately, carried out
It, can quick definitive result figure by the screening of the similarity calculation and structural data of unstructured data when picture search
Picture saves memory space while promoting search efficiency.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to needed in the embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are some embodiments of the invention, general for this field
For logical technical staff, without creative efforts, it is also possible to obtain other drawings based on these drawings.
Fig. 1 is a kind of schematic diagram of the method for picture search provided in an embodiment of the present invention；
Fig. 2 is the schematic diagram of the method for another picture search that the embodiment of the present invention provides；
Fig. 3 is the schematic diagram of the method for another picture search provided in an embodiment of the present invention；
Fig. 4 is a kind of building-block of logic of image search apparatus provided in an embodiment of the present invention；
Fig. 5 provides the building-block of logic of another image search apparatus for the embodiment of the present invention；
Fig. 6 provides the building-block of logic of another image search apparatus for the embodiment of the present invention；
Fig. 7 is a kind of physical structure schematic diagram of image search apparatus provided in an embodiment of the present invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical solution in the embodiment of the present invention are explicitly described, it is clear that described embodiment is the present invention one
The embodiment divided, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not doing
Every other embodiment obtained under the premise of creative work out, should fall within the scope of the present invention.
The term " first " that occurs in description of the invention, claims and attached drawing, " second " and " third " etc. are to use
In the different object of difference, and it is not intended to describe specific sequence.In addition, term " includes " and " having " and they are any
Deformation, it is intended that cover and non-exclusive include.Such as contain the process, method, system, product of a series of steps or units
Or equipment is not limited to listed step or unit, but optionally further comprising the step of not listing or unit, or can
Selection of land further includes the other step or units intrinsic for these process, methods, product or equipment.
Referring to Fig. 1, Fig. 1 is a kind of flow diagram of the method for picture search provided by one embodiment of the present invention.
Wherein, as shown in Figure 1, a kind of method for picture search that one embodiment of the present of invention provides, the method includes following interior
Hold:
101, the inquiry request of user is received, wherein including target image and screening conditions in the inquiry request；
Wherein, executing subject of the invention can be server, this server has search engine functionality.Specifically it can wrap
It includes the various handheld devices with search engine functionality, mobile unit, wearable device, calculate equipment or other processing equipments,
And various forms of user equipmenies (User Equipment, UE), mobile station (Mobile Station, MS), terminal device
(terminal device) etc..Operating system involved by the embodiment of the present application is managed collectively to hardware resource,
And provide a user the software systems of business interface.
Wherein it is possible to understand, the inquiry request of user can request for picture query, or text or language
Sound inquiry request.
For example, which is facial image, and screening conditions can be age, gender, whether with eyes, picture
The parameters such as the time of storage.
In addition, it is necessary to, it is noted that it is described receive user inquiry request before, can also batch some data of increase,
Specifically, the described method includes:
When receiving batch and increasing the request of image, the structural data of the image of the batch and unstructured is obtained
Data；The structural data of the image of the batch is stored into the destination server；And by the image of the batch
Unstructured data store into the preset model.Wherein, the unstructured number of the image data by the batch
According to storage into the preset model, comprising: generate file destination according to the unstructured data of the image of the batch；It uses
The preset model loads the file destination.It is understood that the unstructured data of the image of batch is first generated mesh
File is marked, is the convenience for subsequent load, does not have to thus load unstructured data one by one, to promote processing
Efficiency.
For example, destination server can be solr server (Sol server), which can be faiss
Model (Fa Yisi model).Wherein, solr is an independent search application server, user can by http request, to
XML (extensible markup language) file that application server submits certain format is searched for, index is generated；Http can also be passed through
Get (hypertext transfer protocol acquisition) operation proposes search request, and obtains returning the result for XML format.Faiss model be for
Dense vector provides the frame of efficient similarity search and cluster.Faiss model can provide a variety of retrievals and fast speed.
Specifically, the structural data of image and unstructured data are added to different when batch adds image
In list, wherein first list structured data, second list store unstructured data.Structural data includes figure
The parameters such as time of the gender of personage, age and image storage as in.Unstructured data is the characteristic value of facial image
Data.Batch data in first list is added in solr；By the unstructured data and the image in second list
Mark be added to hdf5 (hdf5 be a kind of file, can handle more objects, the bigger file of storage, support Parallel I/O
(input/output), thread and has other characteristics required by modern operating system and application program, and becomes data model
It is simpler, generality is stronger) in file.Only there are two types of basic structures by hdf5: group (group) and data set (dataset).Group
Include 0 or multiple document data sets.When generating hdf5 file, imitated to reduce to generate the number of file and improve load
Data volume is generated a file according to 1,000,000 secretaries and carries out data storage by rate.In addition, generated using the load of faiss model
Hdf5 file generates the process of central point according to (clustering algorithm) clustering of K-means twice, can all produce to every record
A raw pq code, and be ranked up based on central point.For example, first time k-means is that data are generated one 1024
The central point of cluster calculation is divided into 1024 parts relative to by data.Second of K-means is carried out again to 1024 parts of data
Primary cluster operation.It is understood that second of K-means is each characteristics of image Value Data to be divided into 4 sections, and carrying out
K-means generates 4 cluster points, and each image feature value can generate 4 character string compositions pg code (i.e. four dimensional vectors).
It is understood that use when being to search by the mark storage of image to hdf5.The mark of the image is marked
It is denoted as first identifier.In addition, it is necessary to, it is noted that solr can be every record one document id (file identification) of addition,
The document id is marked as second identifier.It is understood that document id can be a sequence number.
102, the characteristic value of the target image is obtained, and the characteristic value of the target image is input in preset model
To obtain the first identifier of the data of N before similarity ranking；Wherein, N is positive integer；
It is also quickly to search central point (twice according to K-means twice when data search wherein it is possible to understand
Lookup is carried out according to sequencing, that is, is clustered again in the result of first time cluster, then according to secondary cluster
As a result quickly searched), to obtain the highest top n data of similarity, search efficiency can be improved in this way.
103, second identifier corresponding with the first identifier is determined；
Wherein, it should be pointed out that determination second identifier corresponding with the first identifier, comprising: by described
One mark and the mapping table prestored are matched to obtain the second identifier corresponding with the first identifier；Wherein, described
Mapping table is stored with the mapping relations of the first identifier and second identifier, and the first identifier is the figure stored in preset model
The mark of picture, the second identifier are sequence identifier of the described image in the destination server.
104, the second identifier is input in destination server to obtain first structure data；
It is understood that it is corresponding with the second identifier in order to obtain that the second identifier, which is input in solr,
Structural data, i.e. first structure data.
105, the first structure data are screened to obtain the second structural data according to the screening conditions；
It wherein include that third identifies in second structural data, the third is identified as some or all of second identifier；
For example, for example the data of first structure are 100, and screening conditions are male, after screening, are met
The structural data (i.e. the second structural data) of condition is 50.The third that is identified as of so this 50 structural datas identifies
(it is understood that third mark is a logo collection, the inside contains 50 marks).
106, result images are obtained according to third mark and the preset model, and returns to the knot to the user
Fruit image.
It is understood that the third mark be recorded in solr server mark (solr server be every data
The sequence number of editor), it is also necessary to image identification is converted into obtain image corresponding with image identification from faiss model
(the corresponding sequence number of each image is stored in fassi model).
Specifically, described obtain result images according to third mark and the preset model, comprising:
Determine it is corresponding with third mark the 4th identify, wherein the described 4th be identified as the part of first identifier or
All；4th mark is input in the preset model to obtain result images.
For example, since image has the mark (sequence number) of oneself, that is to say, that each image in model has certainly
Oneself sequence number, and in solr server when storing data, one mark (sequence number) can be set for each data.Specifically,
If first identifier (i.e. the sequence number of image in model) is 1,2,3 and 4.It is closed according to the mapping of first identifier and second identifier
System, for example X=10*Y, Y are first identifier, X is second identifier, then corresponding X is 10,20,30 and 40.If being taken in solr
After screening in business device according to screening conditions, as a result only there are two the data fit identified requirements, such as the number that number is 10 and 20
According to meeting screening conditions (i.e. 10 and 20 identify for third), due to 10 and 20 for structural data in solr server mark,
Therefore need to obtain the mark of image in model, according to mapping relations, image is identified as 1 and 2 (the i.e. the 4th marks) in model,
It so obtains in model and identifies 1 and 2 pair of hard image with regard to OK.
As can be seen that the inquiry request of user is received, wherein in the inquiry request in the scheme of the embodiment of the present invention
Including target image and screening conditions；The characteristic value of the target image is obtained, and the characteristic value of the target image is inputted
The first identifier of the data of N before similarity ranking is obtained into preset model；Determine corresponding with the first identifier second
Mark；The second identifier is input in destination server to obtain first structure data；According to the screening conditions pair
The first structure data are screened to obtain the second structural data；It wherein include the in second structural data
Three marks obtain result images according to third mark and the preset model, and return to the result figure to the user
Picture.The structural data of image and unstructured data are stored separately by the technical solution provided through the invention, into
It, can quick definitive result by the screening of the similarity calculation and structural data of unstructured data when row picture search
Image saves memory space while promoting search efficiency.
Referring to Fig. 2, Fig. 2 be another embodiment of the present invention provides another picture search method flow signal
Figure.Wherein, as shown in Figure 2, which comprises
201, when receiving batch and increasing the request of image, the structural data of the image of batch and unstructured is obtained
Data；
202, the structural data of the image of the batch is stored into the destination server；And by the batch
Image unstructured data store into the preset model.
Wherein, the unstructured data of the image data by the batch is stored into the preset model, comprising:
File destination is generated according to the unstructured data of the image of the batch；The target text is loaded using the preset model
Part.
203, the inquiry request of user is received, wherein including target image and screening conditions in the inquiry request；
204, the characteristic value of the target image is obtained, and the characteristic value of the target image is input in preset model
To obtain the first identifier of the data of N before similarity ranking；Wherein, N is positive integer；
205, second identifier corresponding with the first identifier is determined；
Wherein, determination second identifier corresponding with the first identifier, comprising: by the first identifier with prestore
Mapping table is matched to obtain the second identifier corresponding with the first identifier；Wherein, the mapping table table is stored with
The mapping relations of the first identifier and second identifier, the first identifier are the mark of the image stored in preset model, institute
Stating second identifier is sequence identifier of the described image in the destination server.
206, the second identifier is input in destination server to obtain first structure data；
207, the first structure data are screened to obtain the second structural data according to the screening conditions；
It wherein include that third identifies in second structural data, the third is identified as some or all of second identifier；
208, result images are obtained according to third mark and the preset model, and returns to the knot to the user
Fruit image.
It is wherein, described that result images are obtained according to third mark and the preset model, comprising:
Determine it is corresponding with third mark the 4th identify, wherein the described 4th be identified as the part of first identifier or
All；4th mark is input in the preset model to obtain result images.
Wherein, it should be pointed out that the particular content of Fig. 2 described embodiment can refer to embodiment corresponding to Fig. 1
Explanation.
As can be seen that in the scheme of the present embodiment, by the structural data (solr storage) of image of batch addition and non-
Structural data (storage of faiss model) is stored respectively, is stored unstructured data using faiss model, be can be improved
The matched data of image feature data can be improved breakneck acceleration by solr storage organization data and save memory space.It is logical
Cross the search efficiency for further ensuring using technical solution provided in an embodiment of the present invention and promoting user.
As shown in figure 3, another embodiment of the present invention provides another picture search method flow schematic diagram.Its
In, as shown in Figure 3, which comprises
301, when receiving the request of batch increase image, the structural data and non-knot of the image of the batch are obtained
Structure data；
302, the structural data of the image of the batch is stored into the destination server, and obtains the target
Service the first identifier of the structural data of editor；
303, file destination is generated according to the unstructured data of the image of the batch；It is loaded using the preset model
The file destination, and obtain the second identifier of image；
304, the mapping relations of the first identifier and second identifier are established.
305, the inquiry request of user is received, wherein including target image and screening conditions in the inquiry request；
306, the characteristic value of the target image is obtained, and the characteristic value of the target image is input in preset model
To obtain the first identifier of the data of N before similarity ranking；Wherein, N is positive integer；
307, it determines second identifier corresponding with the first identifier, and the second identifier is input to destination service
To obtain first structure data in device；
Wherein, determination second identifier corresponding with the first identifier, comprising: by the first identifier with prestore
Mapping table is matched to obtain the second identifier corresponding with the first identifier；Wherein, the mapping table table is stored with
The mapping relations of the first identifier and second identifier, the first identifier are the mark of the image stored in preset model, institute
Stating second identifier is sequence identifier of the described image in the destination server.
308, the first structure data are screened to obtain the second structural data according to the screening conditions；
It wherein, include that third identifies in second structural data, the third is identified as some or all of second identifier；
309, result images are obtained according to third mark and the preset model, and returns to the knot to the user
Fruit image.
It is wherein, described that result images are obtained according to third mark and the preset model, comprising:
Determine it is corresponding with third mark the 4th identify, wherein the described 4th be identified as the part of first identifier or
All；4th mark is input in the preset model to obtain result images.
Wherein, it should be pointed out that the particular content of Fig. 3 described embodiment can refer to implementation corresponding to Fig. 1 or 2
The explanation of example.
As can be seen that distinguishing in the scheme of the present embodiment by the structural data of image and unstructured data
When storage, the mapping relations of the mark of two storage systems can be established, to can establish the characteristic of image in inquiry
The incidence relation of matching degree and the structural data screening of value.By using technical solution provided in an embodiment of the present invention, into
One step ensure that the search efficiency for promoting user.
As shown in figure 4, a kind of data processing equipment 400 that one embodiment of the present of invention provides, wherein the device 400
Including with lower unit:
Receiving unit 401, for receiving the inquiry request of user, wherein including target image and sieve in the inquiry request
Select condition；
Acquiring unit 402, for obtaining the characteristic value of the target image；
Input unit 403, for the characteristic value of the target image to be input in preset model to obtain similarity row
The first identifier of the data of N before name；Wherein, N is positive integer；
Determination unit 404, for determining second identifier corresponding with the first identifier；
Input unit 403, for the second identifier being input in destination server to obtain first structure data；
Screening unit 405, for being screened according to the screening conditions to the first structure data to obtain
Two structural datas；It wherein include that third identifies in second structural data, the third is identified as the portion of second identifier
Divide or whole；
Acquiring unit 406, for obtaining result images according to third mark and the preset model；
Return unit 407, for returning to the result images to the user.
Optionally, acquiring unit 406 are specifically used for determining the 4th mark corresponding with third mark, wherein described
4th is identified as some or all of first identifier；4th mark is input in the preset model to obtain result figure
Picture.
Optionally, device 400 further includes the first storage unit 408 and the second storage unit 409；
Acquiring unit 406 is also used to obtain the knot of the image of the batch when receiving the request of batch increase image
Structure data and unstructured data；
First storage unit 408, for storing the structural data of the image of the batch to the destination server
In；
Second storage unit 409, for storing the unstructured data of the image of the batch to the preset model
In.
Wherein, the second storage unit 409, the unstructured data for the image according to the batch generate target text
Part；The file destination is loaded using the preset model.
Optionally, determination unit 404, specifically for matching the first identifier with the mapping table prestored to obtain
The second identifier corresponding with the first identifier；Wherein, the mapping table table is stored with the first identifier and the second mark
The mapping relations of knowledge, the first identifier are the mark of the image stored in preset model, and the second identifier is described image
Sequence identifier in the destination server.
Wherein, said units 401-409 can be used for executing method described in step 101-106 in embodiment 1, specifically retouch
Description of the detailed in Example 1 to the method is stated, details are not described herein.
As shown in figure 5, a kind of data processing equipment 500 that one embodiment of the present of invention provides, wherein the device 500
Including with lower unit:
Acquiring unit 501, for obtaining the structure of the image of the batch when receiving the request of batch increase image
Change data and unstructured data；
Storage unit 502, for storing the structural data of the image of the batch into the destination server；With
And the unstructured data of the image of the batch is stored into the preset model.
Wherein, the unstructured data of the image data by the batch is stored into the preset model, comprising:
File destination is generated according to the unstructured data of the image of the batch；The target text is loaded using the preset model
Part.
Receiving unit 503, for receiving the inquiry request of user, wherein including target image and sieve in the inquiry request
Select condition；
Acquiring unit 504, for obtaining the characteristic value of the target image；
Input unit 505, for the characteristic value of the target image to be input in preset model to obtain similarity row
The first identifier of the data of N before name；Wherein, N is positive integer；
Determination unit 506, for determining second identifier corresponding with the first identifier；
Wherein, determination second identifier corresponding with the first identifier, comprising: by the first identifier with prestore
Mapping table is matched to obtain the second identifier corresponding with the first identifier；Wherein, the mapping table table is stored with
The mapping relations of the first identifier and second identifier, the first identifier are the mark of the image stored in preset model, institute
Stating second identifier is sequence identifier of the described image in the destination server.
Input unit 505, for the second identifier being input in destination server to obtain first structure data；
Screening unit 507, for being screened according to the screening conditions to the first structure data to obtain
Two structural datas；It wherein include that third identifies in second structural data, the third is identified as the portion of second identifier
Divide or whole；
Acquiring unit 501, for obtaining result images according to third mark and the preset model, and to the use
Family returns to the result images.
Wherein, said units 501-507 can be used for executing method described in step 201-208 in embodiment 2, specifically retouch
Description of the detailed in Example 2 to the method is stated, details are not described herein.
As shown in fig. 6, a kind of data processing equipment 600 that one embodiment of the present of invention provides, wherein the device 600
Including with lower unit:
Acquiring unit 601, for obtaining the structure of the image of the batch when receiving the request of batch increase image
Change data and unstructured data；
Storage unit 602, for storing the structural data of the image of the batch into the destination server；
Acquiring unit 601, the first identifier of the structural data for obtaining the destination service editor；
Generation unit 603, the unstructured data for the image according to the batch generate file destination；Using described
Preset model loads the file destination, and obtains the second identifier of image；
Unit 604 is established, for establishing the mapping relations of the first identifier and second identifier.
Receiving unit 605, for receiving the inquiry request of user, wherein including target image and sieve in the inquiry request
Select condition；
Acquiring unit 601 is also used to obtain the characteristic value of the target image；
Input unit 606, for the characteristic value of the target image to be input in preset model to obtain similarity row
The first identifier of the data of N before name；Wherein, N is positive integer；
Determination unit 607 is used for determining second identifier corresponding with the first identifier, and the second identifier is defeated
Enter into destination server to obtain first structure data；
Wherein, determination second identifier corresponding with the first identifier, comprising: by the first identifier with prestore
Mapping table is matched to obtain the second identifier corresponding with the first identifier；Wherein, the mapping table table is stored with
The mapping relations of the first identifier and second identifier, the first identifier are the mark of the image stored in preset model, institute
Stating second identifier is sequence identifier of the described image in the destination server.
Screening unit 608, for being screened according to the screening conditions to the first structure data to obtain
Two structural datas；It wherein include that third identifies in second structural data, the third is identified as the portion of second identifier
Divide or whole；
Acquiring unit 601, for obtaining result images according to third mark and the preset model, and to the use
Family returns to the result images.
Wherein, said units 601-608 can be used for executing method described in step 301-309 in embodiment 2, specifically retouch
Description of the detailed in Example 3 to the method is stated, details are not described herein.
Referring to Fig. 7, in another embodiment of the present invention, providing a kind of data processing equipment 700.Device 700 wraps
Include the hardware such as CPU 701, memory 702, bus 703, transceiver 704.Above-mentioned Fig. 4-logic unit shown in fig. 6 can pass through figure
Hardware device shown in 7 is realized.
Wherein, CPU 701 executes the server program being stored in advance in memory 702, which specifically includes:
The inquiry request of user is received, wherein including target image and screening conditions in the inquiry request；
The characteristic value of the target image is obtained, and the characteristic value of the target image is input in preset model to obtain
Take the first identifier of the data of N before similarity ranking；Wherein, N is positive integer；
Determine second identifier corresponding with the first identifier；
The second identifier is input in destination server to obtain first structure data；
The first structure data are screened to obtain the second structural data according to the screening conditions；Wherein
It include that third identifies in second structural data, the third is identified as some or all of second identifier；
Result images are obtained according to third mark and the preset model, and return to the result figure to the user
Picture.
It is optionally, described that result images are obtained according to third mark and the preset model, comprising:
Determine it is corresponding with third mark the 4th identify, wherein the described 4th be identified as the part of first identifier or
All；
4th mark is input in the preset model to obtain result images.
Optionally, before the inquiry request for receiving user, the implementation procedure further include:
When receiving batch and increasing the request of image, the structural data of the image of the batch and unstructured is obtained
Data；
The structural data of the image of the batch is stored into the destination server；And
The unstructured data of the image of the batch is stored into the preset model.
Optionally, the unstructured data of the image data by the batch is stored into the preset model, packet
It includes:
File destination is generated according to the unstructured data of the image of the batch；
The file destination is loaded using the preset model.
Optionally, determination second identifier corresponding with the first identifier, comprising:
The first identifier and the mapping table prestored are matched to obtain corresponding with the first identifier described the
Two marks；Wherein, the mapping table table is stored with the mapping relations of the first identifier and second identifier, and the first identifier is
The mark of the image stored in preset model, the second identifier are sequence mark of the described image in the destination server
Know.
From the above it can be seen that the inquiry request of user is received in technical solution provided in an embodiment of the present invention, wherein the inquiry
It include target image and screening conditions in request；Obtain the characteristic value of the target image, and by the feature of the target image
Value is input to the first identifier that the data of N before similarity ranking are obtained in preset model；Determination is corresponding with the first identifier
Second identifier；The second identifier is input in destination server to obtain first structure data；According to the screening
Condition screens to obtain the second structural data the first structure data；Wherein in second structural data
It is identified including third, result images is obtained according to third mark and the preset model, and to described in user return
Result images.The technical solution provided through the invention is separately deposited the structural data of image and unstructured data
Storage, can be quick by the screening of the similarity calculation and structural data of unstructured data when carrying out picture search
Definitive result image saves memory space while promoting search efficiency.
In another embodiment of the present invention, a kind of computer program product, the computer program product are disclosed
In include program code；When said program code is run, the method in preceding method embodiment can be performed.
In another embodiment of the present invention, a kind of chip is disclosed, includes program code in the chip；Work as institute
When stating program code and being run, the method in preceding method embodiment can be performed.
In several embodiments provided herein, it should be understood that disclosed device, it can be by another way
It realizes.For example, the apparatus embodiments described above are merely exemplary, such as the division of the unit, it is only a kind of
Logical function partition, there may be another division manner in actual implementation, such as multiple units or components can combine or can
To be integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Coupling, direct-coupling or communication connection can be through some interfaces, the indirect coupling or communication connection of device or unit,
It can be electrical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can for personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or
Part steps.And storage medium above-mentioned includes: that USB flash disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic or disk etc. be various to can store program code
Medium.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although referring to before
Stating embodiment, invention is explained in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
It modifies or replaces, the range for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of method of picture search, which is characterized in that the described method includes:
Receive the inquiry request of user, wherein include target image and screening conditions in the inquiry request；
The characteristic value of the target image is obtained, and the characteristic value of the target image is input in preset model to obtain phase
Like the first identifier of the data of N before degree ranking；Wherein, N is positive integer；
Determine second identifier corresponding with the first identifier；
The second identifier is input in destination server to obtain first structure data；
The first structure data are screened to obtain the second structural data according to the screening conditions；It is wherein described
It include that third identifies in second structural data, the third is identified as some or all of second identifier；
Result images are obtained according to third mark and the preset model, and return to the result images to the user.
2. the method according to claim 1, wherein described obtain according to third mark and the preset model
Take result images, comprising:
Determine the 4th mark corresponding with third mark, wherein the described 4th is identified as some or all of first identifier；
4th mark is input in the preset model to obtain result images.
3. according to the method described in claim 2, it is characterized in that, it is described receive user inquiry request before, the method
Further include:
When receiving the request of batch increase image, the structural data and unstructured data of the image of batch are obtained；
The structural data of the image of the batch is stored into the destination server；And
The unstructured data of the image of the batch is stored into the preset model.
4. according to the method described in claim 3, it is characterized in that, the unstructured number of the image data by the batch
According to storage into the preset model, comprising:
File destination is generated according to the unstructured data of the image of the batch；
The file destination is loaded using the preset model.
5. method according to any one of claims 1 to 4, which is characterized in that the determination is corresponding with the first identifier
Second identifier, comprising:
The first identifier and the mapping table prestored are matched to obtain second mark corresponding with the first identifier
Know；Wherein, the mapping table is stored with the mapping relations of the first identifier and second identifier, and the first identifier is default mould
The mark of the image stored in type, the second identifier are sequence identifier of the described image in the destination server.
6. a kind of device of picture search, which is characterized in that described device includes:
Receiving unit, for receiving the inquiry request of user, wherein including target image and screening conditions in the inquiry request；
Acquiring unit, for obtaining the characteristic value of the target image；
Input unit, for the characteristic value of the target image being input in preset model to obtain N before similarity ranking
The first identifier of data；Wherein, N is positive integer；
Determination unit, for determining second identifier corresponding with the first identifier；
The input unit, for the second identifier being input in destination server to obtain first structure data；
Screening unit, for being screened the first structure data to obtain the second structuring according to the screening conditions
Data；It wherein include that third identifies in second structural data, the third is identified as some or all of second identifier；
Acquiring unit, for obtaining result images according to third mark and the preset model；
Return unit, for returning to the result images to the user.
7. device according to claim 6, which is characterized in that the acquiring unit is specifically used for the determining and third
Identify corresponding 4th mark, wherein the described 4th is identified as some or all of first identifier；By the 4th mark input
Into the preset model to obtain result images.
8. device according to claim 7, which is characterized in that described device further includes the first storage unit and the second storage
Unit；
The acquiring unit is also used to obtain the structuring number of the image of batch when receiving the request of batch increase image
According to and unstructured data；
First storage unit, for storing the structural data of the image of the batch into the destination server；
Second storage unit, for storing the unstructured data of the image of the batch into the preset model.
9. device according to claim 8, feature is certainly, second storage unit, for according to the batch
Image unstructured data generate file destination；The file destination is loaded using the preset model.
10. according to any device of claim 6 to 9, which is characterized in that the determination unit, being specifically used for will be described
First identifier and the mapping table prestored are matched to obtain the second identifier corresponding with the first identifier；Wherein, institute
The mapping relations that mapping table table is stored with the first identifier and second identifier are stated, the first identifier is to store in preset model
Image mark, the second identifier be sequence identifier of the described image in the destination server.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811640366.XA CN109829073B (en) | 2018-12-29 | 2018-12-29 | Image searching method and device |
| PCT/CN2019/121607 WO2020134839A1 (en) | 2018-12-29 | 2019-11-28 | Image searching method and apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811640366.XA CN109829073B (en) | 2018-12-29 | 2018-12-29 | Image searching method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109829073A true CN109829073A (en) | 2019-05-31 |
| CN109829073B CN109829073B (en) | 2020-11-24 |
Family
ID=66861411
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811640366.XA Active CN109829073B (en) | 2018-12-29 | 2018-12-29 | Image searching method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109829073B (en) |
| WO (1) | WO2020134839A1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110287346A (en) * | 2019-06-28 | 2019-09-27 | 深圳云天励飞技术有限公司 | Date storage method, device, server and storage medium |
| CN111061888A (en) * | 2019-11-20 | 2020-04-24 | 北京明略软件系统有限公司 | Image acquisition method and system |
| WO2020134839A1 (en) * | 2018-12-29 | 2020-07-02 | 深圳云天励飞技术有限公司 | Image searching method and apparatus |
| CN111966629A (en) * | 2020-06-28 | 2020-11-20 | 电子科技大学 | Particle simulation data storage method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102523304A (en) * | 2011-12-29 | 2012-06-27 | 北京新媒传信科技有限公司 | Application cloud platform and implementation method thereof |
| CN103116643A (en) * | 2013-02-25 | 2013-05-22 | 江苏物联网研究发展中心 | Hadoop-based intelligent medical data management method |
| CN103425780A (en) * | 2013-08-19 | 2013-12-04 | 曙光信息产业股份有限公司 | Data inquiry method and data inquiry device |
| CN106886553A (en) * | 2016-12-27 | 2017-06-23 | 浙江宇视科技有限公司 | A kind of image search method and server |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9575994B2 (en) * | 2011-02-11 | 2017-02-21 | Siemens Aktiengesellschaft | Methods and devices for data retrieval |
| CN106503121B (en) * | 2016-10-19 | 2019-12-06 | 公安部第三研究所 | structured description method and system for X-ray security inspection image |
| CN107346435A (en) * | 2017-06-15 | 2017-11-14 | 浙江捷尚视觉科技股份有限公司 | A kind of suspicion fake-licensed car catching method based on vehicle characteristics storehouse |
| CN107729502A (en) * | 2017-10-18 | 2018-02-23 | 公安部第三研究所 | A kind of bayonet vehicle individualized feature intelligent retrieval system and method |
| CN107944017B (en) * | 2017-12-11 | 2021-06-25 | 浙江捷尚视觉科技股份有限公司 | Method for searching non-motor vehicle in video |
| CN109829073B (en) * | 2018-12-29 | 2020-11-24 | 深圳云天励飞技术有限公司 | Image searching method and device |
-
2018
- 2018-12-29 CN CN201811640366.XA patent/CN109829073B/en active Active
-
2019
- 2019-11-28 WO PCT/CN2019/121607 patent/WO2020134839A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102523304A (en) * | 2011-12-29 | 2012-06-27 | 北京新媒传信科技有限公司 | Application cloud platform and implementation method thereof |
| CN103116643A (en) * | 2013-02-25 | 2013-05-22 | 江苏物联网研究发展中心 | Hadoop-based intelligent medical data management method |
| CN103425780A (en) * | 2013-08-19 | 2013-12-04 | 曙光信息产业股份有限公司 | Data inquiry method and data inquiry device |
| CN106886553A (en) * | 2016-12-27 | 2017-06-23 | 浙江宇视科技有限公司 | A kind of image search method and server |
Non-Patent Citations (2)
| Title |
|---|
| 何龙: ""农业物联网数据存储管理系统的设计与实现"", 《万方》 * |
| 冯喆: ""基于手机的分布式图片检索技术研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020134839A1 (en) * | 2018-12-29 | 2020-07-02 | 深圳云天励飞技术有限公司 | Image searching method and apparatus |
| CN110287346A (en) * | 2019-06-28 | 2019-09-27 | 深圳云天励飞技术有限公司 | Date storage method, device, server and storage medium |
| CN110287346B (en) * | 2019-06-28 | 2021-11-30 | 深圳云天励飞技术有限公司 | Data storage method, device, server and storage medium |
| CN111061888A (en) * | 2019-11-20 | 2020-04-24 | 北京明略软件系统有限公司 | Image acquisition method and system |
| CN111061888B (en) * | 2019-11-20 | 2023-05-16 | 北京明略软件系统有限公司 | Image acquisition method and system |
| CN111966629A (en) * | 2020-06-28 | 2020-11-20 | 电子科技大学 | Particle simulation data storage method |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020134839A1 (en) | 2020-07-02 |
| CN109829073B (en) | 2020-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829073A (en) | A kind of method and device of picture search | |
| US9922051B2 (en) | Image-based faceted system and method | |
| US9400835B2 (en) | Weighting metric for visual search of entity-relationship databases | |
| CN110413875A (en) | A kind of method and relevant apparatus of text information push | |
| CN108288208B (en) | Display object determination method, device, medium and equipment based on image content | |
| CN107147560A (en) | Data sharing method and device | |
| EP2038775A1 (en) | Visual and multi-dimensional search | |
| CN108108743A (en) | Abnormal user recognition methods and the device for identifying abnormal user | |
| JP6865763B2 (en) | Data processing method and equipment | |
| US20230083591A1 (en) | Vector-space framework for evaluating gameplay content in a game environment | |
| WO2019011936A1 (en) | Method for evaluating an image | |
| CN113391992B (en) | Test data generation method and device, storage medium and electronic equipment | |
| CN110166522A (en) | Server identification method, device, readable storage medium storing program for executing and computer equipment | |
| CN109241299A (en) | Multimedia resource searching method, device, storage medium and equipment | |
| CN109885651A (en) | A kind of question pushing method and device | |
| CN109783678B (en) | Image searching method and device | |
| CN110661693A (en) | Methods, computing device-readable storage media, and computing devices facilitating media-based content sharing performed in a computing device | |
| CN104951534B (en) | Search result optimization method and search engine | |
| US10397658B1 (en) | Identifying relevant gameplay content for a game environment using a vector-space framework | |
| CN113590898A (en) | Data retrieval method and device, electronic equipment, storage medium and computer product | |
| CN109961163A (en) | Gender prediction's method, apparatus, storage medium and electronic equipment | |
| CN108491423B (en) | Sorting method and device | |
| CN109815351A (en) | Information query method and Related product | |
| CN106357468B (en) | A kind of method and apparatus for the service regulation judging network server | |
| CN112417290A (en) | Training method of book sorting push model, electronic equipment and 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |