CN106202189A - A kind of image search method and device - Google Patents
A kind of image search method and device Download PDFInfo
- Publication number
- CN106202189A CN106202189A CN201610482644.8A CN201610482644A CN106202189A CN 106202189 A CN106202189 A CN 106202189A CN 201610482644 A CN201610482644 A CN 201610482644A CN 106202189 A CN106202189 A CN 106202189A
- Authority
- CN
- China
- Prior art keywords
- image
- sketch
- matching value
- spatial relation
- searched
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- 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
- G06F16/5854—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using shape and object relationship
Landscapes
- Engineering & Computer Science (AREA)
- Library & Information Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present invention provides a kind of image search method and device, and the lookup method solving browsing pictures one by one in prior art is wasted time and energy, inefficient problem, improves inquiry velocity and the search efficiency of image.Described image search method, including: at picture search interface, receive the sketch of the image to be searched of user terminal input, and obtain the spatial relation between each object that described sketch comprises；According to the spatial relation between each object that the spatial relation between each object that described sketch comprises and the image in image data base comprise, described sketch is mated with the image in image data base, obtain the matching value of image, in the matching value of described image, determine object matching value with preset rules；Image corresponding to described object matching value is defined as the image searched, wherein, described image data base stores multiple image, and the corresponding spatial relation stored between each object that image comprises.
Description
Technical field
The present embodiments relate to technical field of image processing, particularly relate to a kind of image search method and device.
Background technology
Picture library (or photograph album, photo) is system-level application important in mobile terminal, and the picture number in picture library can be along with
It is more and more that user uses the increase of time and accumulates.
In the picture library comprising a large amount of picture, if user wants to search for the picture comprising certain content, in existing method
In, user needs to search, according to shooting time, spot for photography, the picture comprising certain content in the way of browsing one by one.This kind is looked into
Look for mode, situation less for picture number in picture library, lookup demand can be met, but, inventor is realizing the present invention
During find, if the quantity of picture is relatively big in picture library, not only waste time and energy by the way of browsing one by one, and efficiency
Relatively low.
In sum, when in picture library, picture number is more, in prior art, the lookup method of browsing pictures is not only one by one
Waste time and energy, and inefficient.
Summary of the invention
The embodiment of the present invention provides a kind of image search method and device, solves in prior art looking into of browsing pictures one by one
Look for method to waste time and energy, inefficient problem, improve the inquiry velocity of image and search efficiency.
The embodiment of the present invention provides a kind of image search method, and the method includes: at picture search interface, receives user eventually
The sketch of the image to be searched of end input, and obtain the spatial relation between each object that described sketch comprises；According to institute
State the spatial relation between each object that sketch comprises and sky between each object that the image in image data base comprises
Between position relationship, described sketch is mated with the image in image data base, obtains the matching value of image, with preset rules
Object matching value is determined in the matching value of described image；It is defined as the image corresponding to described object matching value searching
Image, wherein, stores multiple image, and the corresponding space bit stored between each object that image comprises in described image data base
Put relation.
A kind of image search apparatus that the embodiment of the present invention provides, this device includes: acquiring unit, in picture search
Interface, receives the sketch of the image to be searched of user terminal input, and obtains the space between each object that described sketch comprises
Position relationship；First processing unit, the spatial relation between each object comprised according to described sketch and picture number
According to the spatial relation between each object that the image in storehouse comprises, described sketch is carried out with the image in image data base
Coupling, obtains the matching value of image, determines object matching value with preset rules in the matching value of described image；Second processes list
Unit, for the image corresponding to described object matching value is defined as the image searched, wherein, deposits in described image data base
Store up multiple image, and the corresponding spatial relation stored between each object that image comprises.
A kind of image search method of embodiment of the present invention offer and device, at picture search interface, receive user terminal
The sketch of the image to be searched of input, and obtain the spatial relation between each object that sketch comprises, comprise according to sketch
Each object between spatial relation and each object of comprising of image in image data base between spatial relation,
Sketch is mated with the image in image data base, obtains the matching value of image, with preset rules at the matching value of image
In determine object matching value, the image corresponding to object matching value is defined as the image searched, namely wraps according in image
Spatial relation between each object contained, comprises the image of certain content, it is not necessary to browse one by one in search image data base
Picture, thus the lookup method that improve browsing pictures one by one in search speed and search efficiency, with prior art is wasted time and energy,
Inefficient comparing, the embodiment of the present invention improves speed and the efficiency of search graph picture of search graph picture.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
In having technology to describe, the required accompanying drawing used is briefly described, it should be apparent that, the accompanying drawing in describing below is this
Some bright embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to root
Other accompanying drawing is obtained according to these accompanying drawings.
The schematic flow diagram of a kind of image search method that Fig. 1 provides for the embodiment of the present invention；
The schematic flow diagram of the idiographic flow of a kind of image search method that Fig. 2 provides for the embodiment of the present invention；
The structural representation of a kind of image search apparatus that Fig. 3 provides for the embodiment of the present invention.
Detailed description of the invention
For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
The a part of embodiment of the present invention rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under not making creative work premise, broadly falls into the scope of protection of the invention.
A kind of image search method that the embodiment of the present invention provides, as it is shown in figure 1, the method includes:
Step 11, at picture search interface, receives the sketch of the image to be searched of user terminal input, and obtains sketch bag
Spatial relation between each object contained；
Wherein, can be picture library (or photograph album, photo) be sent by user terminal receiving user at picture search interface
Picture search instruction after the interface that presents, it is also possible to be that user is when checking picture library (or photograph album, photo) in the user terminal
The interface directly presented, user terminal can be mobile phone, panel computer, camera or wearable smart machine, as the most concrete
Embodiment, as a example by presenting picture search interface after receiving the picture search instruction that user sent by user terminal,
Picture search button is increased, if user is clicked on by user terminal or this image of touch-control in picture library (or photograph album, photo) interface
Search button, then be considered as receiving the picture search instruction that user terminal sends, picture library (or photograph album, photo) display picture search
Interface, prompting user terminal inputs the sketch of image to be searched.
After receiving the sketch of image to be searched of user terminal input, obtain the sky between each object that sketch comprises
Between position relationship, specifically: obtain the spatial relation between each object of comprising of sketch, including: receive user terminal
Spatial relation between each object comprised in the sketch of input；Or the image to be searched according to user terminal input
Sketch, determines the spatial relation between each object comprised in sketch.
When being embodied as, user inputs the sketch of image to be searched by user terminal at picture search interface, if waiting to search
The sketch of rope image is easily drawn, such as: the less and each object detail part of the object comprised in image to be searched is less, then use
The sketch of image to be searched only can be drawn in family at picture search interface, and user terminal is receiving user at picture search interface
After the sketch of the image to be searched of input, according to the sketch of the image to be searched of input, identify in the sketch of image to be searched
The object comprised, and then determine the spatial relation between each object comprised in sketch；If the sketch of image to be searched is difficult
To draw or partly to be difficult to draw in the sketch of image to be searched, then user draws image to be searched at picture search interface
After sketch, it is also possible to use word the part being difficult in sketch draw to be described, such as: use word to describe to be searched
Spatial relation between each object in image, namely user terminal is treating of receiving that user inputs at picture search interface
After the sketch of search graph picture, receive user locus in the sketch that picture search interface inputs between each object and close
System.Certainly, the most preferably, on the basis of picture search interface inputs the sketch of image to be searched, then input in sketch each
Spatial relation between object, to improve the accuracy of search.
If it should be noted that the position that user does not inputs at picture search interface in sketch to be searched between each object is closed
System, then user terminal is after receiving the sketch of image to be searched of user's input, according to the image to be searched of user's input
Sketch, determine the spatial relation between each object comprised in sketch, including: use the degree of depth object detection network detection
The each object comprised in sketch；Use the spatial relation between each object in degree of depth event detection network detection sketch.
When being embodied as, use degree of depth object detection network to detect each object comprised in the sketch of image to be searched, adopt
Detect the spatial relation between each object in the sketch of image to be searched with degree of depth event detection network, specifically include following
Step:
(1) the predefined object classification needing detection, such as: people, plant, animal, building, automobile etc., by degree of depth thing
Health check-up survey grid network (such as: Fast-RCNN), is trained the detector of each type objects；
(2) detection-phase, the sketch first treating search graph picture carries out over-segmentation process, each piece to over-segmentation image
Seek bounding box, it is thus achieved that detection candidate region；
(3) for each detection candidate region, this detection candidate region is input to the object detection network (example trained
As: Fast-RCNN), it is thus achieved that the object detection result in this detection candidate region；
(4) set up the spatial relation between object, specifically, represent the object detected by graph data structure,
Such as: the object that image detects includes: people, animal and plant, then set up an image comprising 3 summits, root
According to object appearance position in the picture, setting up the connection limit between two two articles, the numerical value on limit takes one in 0,1,2,3
Value, each value represents position relationship up and down, such as: 0 represent upper, 1 represent under, 2 represents a left side, 3 represent the right side.
Step 13, comprises with the image in image data base according to the spatial relation between each object that sketch comprises
Each object between spatial relation, sketch is mated with the image in image data base, obtains the coupling of image
Value, determines object matching value with preset rules in the matching value of image；
When being embodied as, according to the spatial relation between each object comprised in image, by the grass of image to be searched
Figure mates one by one with the multiple images comprised in image data base, obtains the matching value of multiple image, and with preset rules
Object matching value is determined, for the ease of according to the sky between object each in image in the matching value of calculated multiple images
Between position relationship, the sketch of image to be searched is mated one by one with the multiple images comprised in image data base, picture number
According to storehouse not only stores the view data of multiple image, but also between each object comprised in every image of corresponding storage
Spatial relation, the determination method of the spatial relation between each object that in image data base, every image comprises, with
Determine that the method for spatial relation between each object comprised in the sketch of image to be searched is identical, namely for picture number
According to every image in storehouse, use each object comprised in degree of depth object detection network detection image, use degree of depth event detection
Network detects the spatial relation in image between each object, and here is omitted.
What deserves to be explained is, according to the spatial relation between each object comprised in image, by image to be searched
Sketch mates one by one with the multiple images comprised in image data base, when being embodied as, by the sketch of image to be searched
Space bit between spatial relation between each object comprised, and each object that in image data base, each image comprises
The relation of putting is mated, and the algorithm of coupling can use algorithm of the prior art, and here is omitted.As the most concrete
Embodiment, it is assumed that comprise people and Ma in the sketch of the image to be searched that user terminal inputs at picture search interface, and user is eventually
Spatial relation between each object comprised in the sketch of end input is: people is on the horse left side, then when mating, according to
Spatial relation between each object, i.e. people are on the horse left side, by comprise in the sketch of image to be searched and image data base
Multiple images mate one by one.
What deserves to be explained is, after obtaining the matching value of multiple image, with preset rules at the matching value of multiple images
In determine object matching value, including: the matching value of multiple images will be defined as mesh more than or equal to the matching value of predetermined threshold value
Marking matching value, namely preset rules is to choose the matching value more than or equal to predetermined threshold value, wherein, predetermined threshold value can freely set
Fixed, such as: predetermined threshold value is 85.Certainly, in other embodiments of the present invention, preset rules can also have other embodiment,
Such as: choose maximum in multiple matching value as object matching value, choose in multiple matching value bigger N number of as target
Matching value, N is natural number.
Step 15, is defined as the image corresponding to object matching value the image searched, wherein, deposits in image data base
Store up multiple image, and the corresponding spatial relation stored between each object that image comprises.
When being embodied as, the image corresponding to object matching value is defined as the image searched, specifically, calculates mesh
The image of mark matching value includes sketch and the picture number of the image to be searched that two width images, i.e. user are inputted by user terminal
According to an image in storehouse, then the image corresponding to object matching value is defined as the image searched, including: by object matching
The image in image data base corresponding to value is defined as the image searched.
As more specific embodiment, it is assumed that determine object matching value in the matching value of multiple images with preset rules
In preset rules be to choose the matching value more than or equal to predetermined threshold value, the value of predetermined threshold value is 90, it is assumed that figure to be searched
Image A in the sketch of picture and image data base, image B, image C, image D, the matching value of image E be respectively 20,70,55,90,
95, then the object matching value determined with preset rules is 90,95, in the image data base corresponding to object matching value 90
Image is image D, and in the image data base corresponding to object matching value 95, image is image E, then determined by image D and image E
For the image searched.
As another more specific embodiment, it is assumed that determine target with preset rules in the matching value of multiple images
Joining the preset rules in value is to choose maximum, it is still assumed that the sketch of image to be searched and image A, image in image data base
B, image C, image D, the matching value of image E are respectively 20,70,55,90,95, then the object matching determined with preset rules
Value is 95, and in the image data base corresponding to object matching value 95, image is image E, then be defined as searching by image E
Image.
As another more specific embodiment, it is assumed that determine target with preset rules in the matching value of multiple images
Join the preset rules in value be bigger in the matching value choosing multiple image 3 as object matching value, it is still assumed that wait to search
Image A in the sketch of rope image and image data base, image B, image C, image D, the matching value of image E be respectively 20,70,
55,90,95, then the object matching value determined with preset rules is 95,90,70, due to the figure corresponding to object matching value 95
As in data base, image is image E, in the image data base corresponding to object matching value 90, image is image D, object matching value
In image data base corresponding to 70, image is image B, then image B, image D and image E are defined as the image searched.
In the method that the embodiment of the present invention provides, at picture search interface, receive the image to be searched of user terminal input
Sketch, and obtain the spatial relation between each object that sketch comprises, according to the sky between each object that sketch comprises
Between spatial relation between each object of comprising of image in position relationship and image data base, by sketch and view data
Image in storehouse mates, and obtains the matching value of image, determines object matching value with preset rules in the matching value of image,
Image corresponding to object matching value is defined as the image searched, namely according to the sky between each object comprised in image
Between position relationship, search image data base comprises the image of certain content, it is not necessary to browsing pictures one by one, thus improves search
In speed and search efficiency, with prior art, the lookup method of browsing pictures is wasted time and energy one by one, inefficient compares, the present invention
Embodiment improves speed and the efficiency of search graph picture of search graph picture.
In a kind of possible embodiment, in the method that the embodiment of the present invention provides, obtain each object that sketch comprises
Between spatial relation, including: the locus received in the sketch of user terminal input between each object of comprising is closed
System；Or the sketch of the image to be searched according to user terminal input, determines the space bit between each object comprised in sketch
Put relation.
In a kind of possible embodiment, in the method that the embodiment of the present invention provides, according to treating of user terminal input
The sketch of search graph picture, determines the spatial relation between each object comprised in sketch, including: use degree of depth object detection
The each object comprised in network detection sketch；Use the locus between each object in degree of depth event detection network detection sketch
Relation.
In a kind of possible embodiment, in the method that the embodiment of the present invention provides, the method also includes: search at image
Rope interface pre-sets the image of the object that multiple images in image data base comprise.
When being embodied as, the embodiment of the present invention, when carrying out picture search, needs in the input of picture search interface to be searched
The sketch of image, by pre-setting the image of the object that multiple images in image data base comprise at picture search interface,
User terminal can be facilitated to input the sketch of image to be searched at search interface, facilitate the image that user terminal selecting pre-sets
Complete the drafting of sketch at picture search interface, use the image rendering sketch pre-set simultaneously, improve the standard of skeletonizing
Really property, further, by pre-setting the figure of the object that multiple images in image data base comprise at picture search interface
Picture so that the image that user terminal selecting pre-sets completes the drafting of sketch at picture search interface, can improve sketch
In the recognition accuracy of object that comprises, it is accurate to the identification of the spatial relation between each object comprised in sketch to improve
Rate, thus improve the accuracy rate of search graph picture.
In a kind of possible embodiment, the embodiment of the present invention provide method in, with preset rules image
Join and value determines object matching value, including: the matching value being more than or equal to predetermined threshold value in the matching value of image is defined as mesh
Mark matching value.
The image search method provided the embodiment of the present invention below in conjunction with Fig. 2 is described in detail.As in figure 2 it is shown, this
The concrete steps of the image search method that inventive embodiments provides, including:
Step 21, user terminal is opened picture search function, is pointed out user terminal to draw the sketch of image to be searched, specifically
For, user terminal needs to search for the image comprising certain content in picture library, then open the sketch function of search in picture library, according to
Prompting skeletonizing on drawing board, for being difficult to the part drawn, it is possible to use vocabulary describes, and inputs to search module；
Step 23, user terminal carries out pretreatment by built-in algorithm to every image in picture library, identifies every
The spatial relation between object and object in image, specifically, by degree of depth learning method, important in image
Object positions, and position occurs in detection object, and identifies the spatial relation between object, such as: people is sitting in sofa
On；
Step 25, according to the spatial relation between the object comprised in image and object, inputs user terminal
The sketch image each with picture library of image to be searched mate, and determine object matching value according to preset rules, determine
Image corresponding to object matching value, specifically, according to the position relationship between object and the object described in sketch, passes through
The sketch of the image to be searched that user terminal is inputted by matching algorithm mates with each image in picture library, obtains multiple figure
The matching value of picture, and in the matching value of multiple images, determine object matching value with preset rules, such as: select multiple image
3 or 5 images that in matching value, matching degree is higher；
Step 27, presents to user by the image corresponding to object matching value by picture search interface, specifically, and will
Image corresponding to object matching value shows at picture search interface, allows user select, and finds out picture to be searched.
A kind of image search apparatus that the embodiment of the present invention provides, as it is shown on figure 3, this device includes: acquiring unit 31, uses
In at picture search interface, receive the sketch of the image to be searched of user terminal input, and obtain each object that sketch comprises it
Between spatial relation；First processing unit 33, the spatial relation between each object comprised according to sketch with
Spatial relation between each object that image in image data base comprises, enters sketch with the image in image data base
Row coupling, obtains the matching value of image, determines object matching value with preset rules in the matching value of image；Second processing unit
35, for the image corresponding to object matching value being defined as the image searched, wherein, image data base stores multiple figure
Picture, and the corresponding spatial relation stored between each object that image comprises.
In the device that the embodiment of the present invention provides, at picture search interface, receive the image to be searched of user terminal input
Sketch, and obtain the spatial relation between each object that sketch comprises, according to the sky between each object that sketch comprises
Between spatial relation between each object of comprising of image in position relationship and image data base, by sketch and view data
Image in storehouse mates, and obtains the matching value of image, determines object matching value with preset rules in the matching value of image,
Image corresponding to object matching value is defined as the image searched, namely according to the sky between each object comprised in image
Between position relationship, search image data base comprises the image of certain content, it is not necessary to browsing pictures one by one, thus improves search
In speed and search efficiency, with prior art, the lookup method of browsing pictures is wasted time and energy one by one, inefficient compares, the present invention
Embodiment improves speed and the efficiency of search graph picture of search graph picture.
In a kind of possible embodiment, in the device that the embodiment of the present invention provides, acquiring unit 31 obtains sketch bag
Spatial relation between each object contained, specifically for: receive in the sketch of user terminal input each object of comprising it
Between spatial relation；Or the sketch of the image to be searched according to user terminal input, determines each thing comprised in sketch
Spatial relation between body.
In a kind of possible embodiment, in the device that the embodiment of the present invention provides, acquiring unit 31 is according to user eventually
The sketch of the image to be searched of end input, determines the spatial relation between each object comprised in sketch, specifically for: adopt
With each object comprised in degree of depth object detection network detection sketch；Use each object in degree of depth event detection network detection sketch
Between spatial relation.
In a kind of possible embodiment, in the device that the embodiment of the present invention provides, the first processing unit 33, also use
In: the image of the object comprised in picture search interface pre-sets multiple images in image data base.
In a kind of possible embodiment, in the device that the embodiment of the present invention provides, the first processing unit 33 is to preset
Rule determines object matching value in the matching value of image, specifically for: preset threshold by the matching value of image is more than or equal to
The matching value of value is defined as object matching value.
The embodiment of the present invention provide image search apparatus, can integrated in the user terminal, such as: be integrated in mobile phone
In, it is also possible to it is integrated in image processing software, wherein, acquiring unit the 31, first processing unit 33 and the second processing unit 35
All can use the processors such as CPU.
A kind of image search method of embodiment of the present invention offer and device, at picture search interface, receive user terminal
The sketch of the image to be searched of input, and obtain the spatial relation between each object that sketch comprises, comprise according to sketch
Each object between spatial relation and each object of comprising of image in image data base between spatial relation,
Sketch is mated with the image in image data base, obtains the matching value of image, with preset rules at the matching value of image
In determine object matching value, the image corresponding to object matching value is defined as the image searched, namely wraps according in image
Spatial relation between each object contained, comprises the image of certain content, it is not necessary to browse one by one in search image data base
Picture, thus improve speed and the efficiency of search graph picture of search graph picture.
Device embodiment described above is only schematically, and the wherein said unit illustrated as separating component can
To be or to may not be physically separate, the parts shown as unit can be or may not be physics list
Unit, i.e. may be located at a place, or can also be distributed on multiple NE.Can be selected it according to the actual needs
In some or all of module realize the purpose of the present embodiment scheme.Those of ordinary skill in the art are not paying creativeness
Work in the case of, be i.e. appreciated that and implement.
Through the above description of the embodiments, those skilled in the art it can be understood that to each embodiment can
The mode adding required general hardware platform by software realizes, naturally it is also possible to pass through hardware.Based on such understanding, on
State the part that prior art contributes by technical scheme the most in other words to embody with the form of software product, should
Computer software product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD etc., including some fingers
Make with so that a computer equipment (can be personal computer, server, or the network equipment etc.) performs each and implements
The method described in some part of example or embodiment.
Last it is noted that above example is only in order to illustrate technical scheme, it is not intended to limit；Although
With reference to previous embodiment, the present invention is described in detail, it will be understood by those within the art that: it still may be used
So that the technical scheme described in foregoing embodiments to be modified, or wherein portion of techniques feature is carried out equivalent；
And these amendment or replace, do not make appropriate technical solution essence depart from various embodiments of the present invention technical scheme spirit and
Scope.
Claims (10)
1. an image search method, it is characterised in that the method includes:
At picture search interface, receive the sketch of image to be searched of user terminal input, and obtain that described sketch comprises each
Spatial relation between object；
The each thing comprised with the image in image data base according to the spatial relation between each object that described sketch comprises
Spatial relation between body, mates described sketch with the image in image data base, obtains the matching value of image,
In the matching value of described image, object matching value is determined with preset rules；
Image corresponding to described object matching value is defined as the image searched, wherein, described image data base stores
Multiple images, and the corresponding spatial relation stored between each object that image comprises.
Method the most according to claim 1, it is characterised in that the sky between each object that the described sketch of described acquisition comprises
Between position relationship, including:
Receive the spatial relation between each object comprised in the described sketch of user terminal input；Or according to user eventually
The sketch of the image to be searched of end input, determines the spatial relation between each object comprised in described sketch.
Method the most according to claim 2, it is characterised in that according to the sketch of the image to be searched of user terminal input,
Determine the spatial relation between each object comprised in described sketch, including:
Degree of depth object detection network is used to detect each object comprised in described sketch；
Use the spatial relation between each object in the degree of depth event detection network described sketch of detection.
Method the most according to claim 1, it is characterised in that the method also includes:
The image of the object that multiple images in picture search interface pre-sets described image data base comprise.
5. according to the method according to any one of claim 1-4, it is characterised in that with preset rules in the coupling of described image
Value determines object matching value, including:
The matching value being more than or equal to predetermined threshold value in the matching value of described image is defined as object matching value.
6. an image search apparatus, it is characterised in that this device includes:
Acquiring unit, at picture search interface, receives the sketch of the image to be searched of user terminal input, and obtains described
Spatial relation between each object that sketch comprises；
First processing unit, the spatial relation between each object comprised according to described sketch with in image data base
Each object of comprising of image between spatial relation, described sketch is mated with the image in image data base,
Obtain the matching value of image, in the matching value of described image, determine object matching value with preset rules；
Second processing unit, for the image corresponding to described object matching value being defined as the image searched, wherein, described
Image data base stores multiple image, and the corresponding spatial relation stored between each object that image comprises.
Device the most according to claim 6, it is characterised in that described acquiring unit obtains each object that described sketch comprises
Between spatial relation, specifically for:
Receive the spatial relation between each object comprised in the described sketch of user terminal input；Or according to user eventually
The sketch of the image to be searched of end input, determines the spatial relation between each object comprised in described sketch.
Device the most according to claim 7, it is characterised in that described acquiring unit according to user terminal input to be searched
The sketch of image, determines the spatial relation between each object comprised in described sketch, specifically for:
Degree of depth object detection network is used to detect each object comprised in described sketch；
Use the spatial relation between each object in the degree of depth event detection network described sketch of detection.
Device the most according to claim 6, it is characterised in that described first processing unit, is additionally operable to: in picture search circle
Face pre-sets the image of the object comprised in multiple images in described image data base.
10. according to the device according to any one of claim 6-9, it is characterised in that described first processing unit is with default rule
In the matching value of described image, then determine object matching value, specifically for:
The matching value being more than or equal to predetermined threshold value in the matching value of described image is defined as object matching value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610482644.8A CN106202189A (en) | 2016-06-27 | 2016-06-27 | A kind of image search method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610482644.8A CN106202189A (en) | 2016-06-27 | 2016-06-27 | A kind of image search method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106202189A true CN106202189A (en) | 2016-12-07 |
Family
ID=57461896
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610482644.8A Pending CN106202189A (en) | 2016-06-27 | 2016-06-27 | A kind of image search method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106202189A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106814967A (en) * | 2017-01-25 | 2017-06-09 | 努比亚技术有限公司 | The apparatus and method of retrieving image in a kind of picture library |
| CN109635139A (en) * | 2018-11-13 | 2019-04-16 | 北京合享智慧科技有限公司 | A kind of method and relevant apparatus obtaining image information |
| CN110110117A (en) * | 2017-12-20 | 2019-08-09 | 阿里巴巴集团控股有限公司 | A kind of product search method, device and system |
| CN110532415A (en) * | 2019-08-30 | 2019-12-03 | 百度在线网络技术（北京）有限公司 | Picture search processing method, device, equipment and storage medium |
| CN110929057A (en) * | 2018-08-30 | 2020-03-27 | 深圳市蓝灯鱼智能科技有限公司 | Image processing method, device and system, storage medium and electronic device |
| CN111563181A (en) * | 2020-05-12 | 2020-08-21 | 海口科博瑞信息科技有限公司 | Digital image file query method and device and readable storage medium |
| US10789288B1 (en) * | 2018-05-17 | 2020-09-29 | Shutterstock, Inc. | Relational model based natural language querying to identify object relationships in scene |
| CN112069336A (en) * | 2020-08-04 | 2020-12-11 | 中国科学院软件研究所 | Fine-grained image retrieval method and system based on scene sketch |
| CN112307243A (en) * | 2019-07-23 | 2021-02-02 | 北京京东振世信息技术有限公司 | Method and apparatus for retrieving image |
| CN112364199A (en) * | 2021-01-13 | 2021-02-12 | 太极计算机股份有限公司 | Picture searching system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102339306A (en) * | 2010-08-31 | 2012-02-01 | 微软公司 | Sketch-based image search |
| CN104737160A (en) * | 2012-10-15 | 2015-06-24 | 微软公司 | Pictures from sketches |
| CN105493078A (en) * | 2013-06-14 | 2016-04-13 | 微软技术许可有限责任公司 | Color sketch image searching |
-
2016
- 2016-06-27 CN CN201610482644.8A patent/CN106202189A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102339306A (en) * | 2010-08-31 | 2012-02-01 | 微软公司 | Sketch-based image search |
| CN104737160A (en) * | 2012-10-15 | 2015-06-24 | 微软公司 | Pictures from sketches |
| CN105493078A (en) * | 2013-06-14 | 2016-04-13 | 微软技术许可有限责任公司 | Color sketch image searching |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106814967A (en) * | 2017-01-25 | 2017-06-09 | 努比亚技术有限公司 | The apparatus and method of retrieving image in a kind of picture library |
| CN110110117A (en) * | 2017-12-20 | 2019-08-09 | 阿里巴巴集团控股有限公司 | A kind of product search method, device and system |
| US10789288B1 (en) * | 2018-05-17 | 2020-09-29 | Shutterstock, Inc. | Relational model based natural language querying to identify object relationships in scene |
| CN110929057A (en) * | 2018-08-30 | 2020-03-27 | 深圳市蓝灯鱼智能科技有限公司 | Image processing method, device and system, storage medium and electronic device |
| CN109635139A (en) * | 2018-11-13 | 2019-04-16 | 北京合享智慧科技有限公司 | A kind of method and relevant apparatus obtaining image information |
| CN109635139B (en) * | 2018-11-13 | 2020-05-01 | 北京合享智慧科技有限公司 | Method and related device for acquiring image information |
| CN112307243A (en) * | 2019-07-23 | 2021-02-02 | 北京京东振世信息技术有限公司 | Method and apparatus for retrieving image |
| US12106555B2 (en) | 2019-07-23 | 2024-10-01 | Beijing Jingdong Zhenshi Information Technology Co., Ltd. | Method and device for retrieving image |
| CN112307243B (en) * | 2019-07-23 | 2023-11-03 | 北京京东振世信息技术有限公司 | Method and apparatus for retrieving images |
| CN110532415A (en) * | 2019-08-30 | 2019-12-03 | 百度在线网络技术（北京）有限公司 | Picture search processing method, device, equipment and storage medium |
| CN111563181B (en) * | 2020-05-12 | 2023-05-05 | 海口科博瑞信息科技有限公司 | Digital image file query method, device and readable storage medium |
| CN111563181A (en) * | 2020-05-12 | 2020-08-21 | 海口科博瑞信息科技有限公司 | Digital image file query method and device and readable storage medium |
| CN112069336A (en) * | 2020-08-04 | 2020-12-11 | 中国科学院软件研究所 | Fine-grained image retrieval method and system based on scene sketch |
| CN112364199A (en) * | 2021-01-13 | 2021-02-12 | 太极计算机股份有限公司 | Picture searching system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106202189A (en) | A kind of image search method and device | |
| KR102061288B1 (en) | Image retrieval method, real-time drawing prompting method and devices thereof | |
| US11222044B2 (en) | Natural language image search | |
| CN104765891A (en) | Searching shopping method based on pictures | |
| US11036790B1 (en) | Identifying visual portions of visual media files responsive to visual portions of media files submitted as search queries | |
| CN105493078B (en) | Colored sketches picture search | |
| CN104537341B (en) | Face picture information getting method and device | |
| CN110472082B (en) | Data processing method, data processing device, storage medium and electronic equipment | |
| CN111275060B (en) | Identification model updating processing method and device, electronic equipment and storage medium | |
| WO2021208633A1 (en) | Method and device for determining item name, computer apparatus, and storage medium | |
| CN111581423B (en) | Target retrieval method and device | |
| CN106339479A (en) | Picture naming method and terminal | |
| CN110059156A (en) | Coordinate retrieval method, apparatus, equipment and readable storage medium storing program for executing based on conjunctive word | |
| US10740385B1 (en) | Identifying visual portions of visual media files responsive to search queries | |
| JP2007094679A (en) | Image analyzing device, image analyzing program and image analyzing program storage medium | |
| Karaoglu et al. | Con-text: text detection using background connectivity for fine-grained object classification | |
| CN110019913A (en) | Picture match method, user equipment, storage medium and device | |
| US12072925B2 (en) | Visual search via free-form visual feature selection | |
| CN107491521A (en) | A kind of image search system and method | |
| CN111783786B (en) | Picture identification method, system, electronic device and storage medium | |
| Matsui | Challenge for manga processing: Sketch-based manga retrieval | |
| Farhat et al. | Intelligent portrait composition assistance: Integrating deep-learned models and photography idea retrieval | |
| Boteanu et al. | Hierarchical clustering pseudo-relevance feedback for social image search result diversification | |
| CN110472079B (en) | Target image retrieval method, device, equipment and storage medium | |
| Saravanan et al. | Image retrieval by image feature using data mining technique |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20161207 |
|
| WD01 | Invention patent application deemed withdrawn after publication |