CN105975922A - Information processing method and information processing device - Google Patents
Information processing method and information processing device Download PDFInfo
- Publication number
- CN105975922A CN105975922A CN201610281956.2A CN201610281956A CN105975922A CN 105975922 A CN105975922 A CN 105975922A CN 201610281956 A CN201610281956 A CN 201610281956A CN 105975922 A CN105975922 A CN 105975922A
- Authority
- CN
- China
- Prior art keywords
- pending image
- image block
- pending
- information
- block
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention provides an information processing method and an information processing device. According to one embodiment of the invention, the information processing method includes the following steps that: image segmentation is performed on an image to be processed, so that at least one image block to be processed can be obtained; the image blocks to be processed are detected through a category classifier, so that a detection result can be obtained, wherein the detection result contains categories corresponding to the image blocks to be processed; the image blocks to be processed are detected through attribute classifiers corresponding to the determined categories, so that attributes corresponding to the image blocks to be processed can be determined; description information corresponding to the determined attributes are adopted as description information of an object contained in the image to be processed; and when an acquisition request which is used for acquiring the information of the object contained in the image to be processed and is sent by a client is received, the description information is sent to the client, so that the client can display the description information. With the information processing method and device adopted, a user does not need to manually input character information at a web page as what is required in the prior art to search the description information of an object which the user is interested in, and therefore, the efficiency of information acquisition can be improved.
Description
Technical field
The present embodiments relate to technical field of image processing, particularly relate to information processing method and device.
Background technology
Along with the development of network technology, the digital picture feature as user discrimination visual in image with it, easy is filled
Scold network.Such as, when user is when browsing webpage, viewing video, be seen everywhere image.Image wraps
Include at least one object, and user may can be interested be in this object.Such as, when user watches video, can
Can be interested in the clothes of performer, it is desirable to understand this part clothes more attribute informations (such as, seller,
Price, the place of production etc.).
And in prior art, if user wants to understand the attribute information of objects in image, hands can only be passed through
Dynamic acquisition.Such as, user relies on the impression of self to decide the Word message describing objects, then leads to
Cross and input corresponding word to search for objects at webpage, retouch obtaining the more of objects from network
State information.
And the method for the description information of this object for obtaining in image, operation complexity, and inefficiency.
Summary of the invention
The embodiment of the present invention provides information processing method and device, in order to solve at present due to, if user thinks
Solve the attribute information of objects in image, can only manually obtain, cause the letter for obtaining information
The problem of breath processing method, operation complexity, and inefficiency etc..
The concrete technical scheme that the embodiment of the present invention provides is as follows:
On the one hand, the embodiment of the present invention provides a kind of information processing method, and described method includes:
Pending image is carried out image segmentation, it is thus achieved that at least one pending image block；
For this pending image block, by category classifier, this process image block is detected, it is thus achieved that inspection
Surveying result, this testing result includes the classification that this pending image block is corresponding；And,
By the attributive classification device corresponding with the classification determined, this pending image block is detected, determines
The attribute that this pending image block is corresponding；And,
Using description information corresponding for the attribute determined as the description of the object comprised in described pending image
Information；
When the acquisition receiving the information obtaining the object comprised in described pending image that client sends
During request, described description information is sent to client and shows.
On the other hand, the present invention provides a kind of information processor, including:
Segmentation module, for carrying out image segmentation to pending image, it is thus achieved that at least one pending image block；
Detection module, for for this pending image block, is entered this process image block by category classifier
Row detection, it is thus achieved that testing result, this testing result includes the classification that this pending image block is corresponding；
Attribute determination module, for by the attributive classification device corresponding with the classification determined, to this pending figure
As block detects, determine the attribute that this pending image block is corresponding；
Information determination module is described, for by description information corresponding for the attribute that determine as described pending figure
The description information of the object comprised in Xiang；
Sending module, for when receiving the object comprised in the described pending image of acquisition that client sends
Information acquisition request time, described description information is sent to client and shows.
Beneficial effect of the present invention includes: in the embodiment of the present invention, pending image is carried out image segmentation, obtains
Obtain at least one pending image block；For each pending image block, by category classifier to this process
Image block detects, it is thus achieved that testing result, and this testing result includes the class that this pending image block is corresponding
Not；And, by the attributive classification device corresponding with the classification determined, this pending image block is detected,
Determine the attribute that this pending image block is corresponding；And, using description information corresponding for the attribute that determines as described
The description information of the object comprised in pending image；When the acquisition receiving client transmission is described pending
When the acquisition of the information of the object comprised in image is asked, described description information is sent to client and shows.
So, in the embodiment of the present invention, by pending image is analyzed and obtains comprise in this image right
The description information pushing of elephant is to user so that user is without manually inputting word at webpage as prior art
The description information of information search objects, relative to prior art, the embodiment of the present invention can improve letter
The efficiency that breath obtains.
Accompanying drawing explanation
Fig. 1 is the exemplary process diagram of the information processing method in the embodiment of the present invention one；
Fig. 2 is that the destination object in the embodiment of the present invention one has erose schematic diagram；
Fig. 3 is the display description information in the embodiment of the present invention one and the schematic diagram of analogical object；
Fig. 4 is the structural representation of the information processor in the embodiment of the present invention two.
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 present invention
Accompanying drawing in embodiment, is clearly and completely described the technical scheme in the embodiment of the present invention, it is clear that
Described embodiment is a part of embodiment of the present invention rather than whole embodiments.Based in the present invention
Embodiment, those of ordinary skill in the art obtained under not making creative work premise all its
His embodiment, broadly falls into the scope of protection of the invention.
In the embodiment of the present invention, first building category classifier, category grader user identifies pending figure
The classification belonging to object comprised in Xiang.For example, it is possible to category classifier is entered by the sample obtained in advance
Row training obtains category classifier.Such as, pre-defined classification includes jacket, medicated cap, trousers, wraps, OK
Lee's case etc..Then use jacket training sample, medicated cap training sample, trousers training sample, bag training sample,
Category classifier is trained by luggage case training sample, so that category classifier includes being capable of identify that jacket
Jacket grader, be capable of identify that the medicated cap grader of medicated cap, be capable of identify that the trousers grader of trousers, energy
Enough luggage case graders etc. identifying the packet classifier wrapped and being capable of identify that luggage case.So, classification is divided
The classification that the object comprised in pending image is corresponding just can be identified by class device.
Additionally, the embodiment of the present invention builds attributive classification device in advance, as a example by jacket, attributive classification device is entered
Row explanation.Such as, jacket can comprise multiple attribute, such as long sleeves, cotta, money of cultivating one's moral character, loose version,
The attributes such as V neck, crew neck.Attributive classification device is trained so that attributive classification device can by jacket training sample
Identify the attribute of jacket.
In the embodiment of the present invention, pending image is carried out image segmentation, it is thus achieved that at least one pending image
Block；For each pending image block, by category classifier, this process image block is detected, it is thus achieved that
Testing result, this testing result includes the classification that this pending image block is corresponding；And, by with determine
The attributive classification device that classification is corresponding, detects this pending image block, determines this pending image block pair
The attribute answered；And, description information corresponding for the attribute determined is right as comprise in described pending image
The description information of elephant；When receiving the letter obtaining the object comprised in described pending image that client sends
When the acquisition of breath is asked, described description information is sent to client and shows.So, in the embodiment of the present invention,
By pending image being analyzed and obtaining the description information pushing of the object comprised in this image to using
Family so that user is without manually searching for objects at webpage inputting word information as prior art
Description information, relative to prior art, the embodiment of the present invention can improve the efficiency of acquisition of information.
For ease of being further appreciated by, below by embodiment, technical solution of the present invention is described further.
Embodiment one
As it is shown in figure 1, the schematic flow sheet of the information processing method provided for the embodiment of the present invention, the method
Comprise the following steps:
Step 101: pending image is carried out image segmentation, it is thus achieved that at least one pending image block.
Wherein, in one embodiment, the method for image segmentation can determine according to prior art, such as base
Dividing method in region, dividing method based on edge and dividing method based on particular theory etc. are the suitableeest
For the embodiment of the present invention, this is not limited by the present invention.
Step 102: for this pending image block, by category classifier, this process image block is examined
Survey, it is thus achieved that testing result, this testing result includes the classification that this pending image block is corresponding.
Wherein it is possible to perform the operation of step 102 for one or more pending image block.
Step 103: by the attributive classification device corresponding with the classification determined, this pending image block is carried out
Detection, determines the attribute that this pending image block is corresponding.
Step 104: description information corresponding for the attribute determined is right as comprise in described pending image
The description information of elephant.
Step 105: when receiving the letter obtaining the object comprised in described pending image that client sends
When the acquisition of breath is asked, described description information is sent to client and shows.
Wherein, in one embodiment, the described pending image of acquisition that reception client sends comprises
The execution obtaining request of the information of object is unrestricted for opportunity.
Wherein, in one embodiment, in order to fine marks off each object that image includes,
In the embodiment of the present invention, in step 101 (i.e. pending image is carried out image segmentation, it is thus achieved that at least one
Image block), may particularly include following steps:
Step A1: pending image is carried out over-segmentation process, it is thus achieved that at least one destination object.
Wherein, in one embodiment, over-segmentation processing method compares other image partition method, permissible
Pending image is partitioned into more pending image block, it is achieved to finer the dividing of pending image
Cut, so ensure that and a complete object also can be partitioned into different pending piece.The most pending
Image includes a people, is processed by over-segmentation and can obtain the head of this people, upper body and the lower part of the body, from
And the medicated cap on this person, jacket and lower clothing can be obtained.
Wherein, when being embodied as, over-segmentation processing method can determine according to prior art, such as
MeanShift dividing method, watershed segmentation methods etc..
Step A2: for this destination object, determines and can comprise this destination object in pending image
Little preset pattern region is as a pending image block.
Wherein it is possible to perform the operation of step A2 for one or more destination object.
Mostly have irregularly shaped (such as figure as in figure 2 it is shown, use over-segmentation to process the destination object identified
In 2 201 are for having erose destination object).If according to this irregularly shaped as segmentation
Border, then process complexity, and split obtain image may lose destination object partial information (because
Some information may be outside border).So, in the embodiment of the present invention, in order to improve the effect obtaining image block
Rate, for each destination object, determines that the minimum that can comprise this destination object in pending image presets figure
Shape region (such as 202 in Fig. 2) is as a pending image block, so without along irregular limit
Boundary's segmentation image, it is simple to pending image is split and pending image block can be made to include mesh as far as possible
The more information of logo image, in order to the later stage passes through category classifier and attributive classification device to pending figure
The destination object included as block is identified.
Wherein, in one embodiment, preset pattern e.g. rectangle, circle, triangle, hexagon etc.,
When being embodied as, can set according to actual needs, this is not limited by the present invention.
Wherein, in one embodiment, when using the over-segmentation in step A1 to process the pending image of acquisition
After block, it is thus achieved that image block may be less.And the content included by image varies, some is special
The complete information including an object that bigger image block can be complete may be obtained in the case of Shu.So,
In the embodiment of the present invention, identify the accurate of the object that comprises in pending image in order to improve further
Property, it is also possible to the method grown by zone similarity, it is thus achieved that more pending image block, concrete, bag
Include following steps:
Step B1: for the two pending image blocks that position is adjacent in pending image, calculate this adjacent
The aberration of two pending image blocks.
Step B2: if aberration is less than presetting aberration, then these adjacent two pending image blocks are merged, and will
Image block after merging is as newly-increased pending image block, and the image block after merging puts into optimization set
In.
At this moment, the pending image that the pending image block of pending image will include that over-segmentation obtains after processing
Block, and newly-increased pending image block.
Step B3: if aberration is more than or equal to presetting aberration, then these adjacent two pending image blocks are put into
In optimization set.
Step B4: for the pending image block in optimization set, returns and performs at pending image
The two pending image blocks that middle position is adjacent, calculate the step of the aberration of these adjacent two pending image blocks.
Wherein, in one embodiment, can be to perform for one or more pending image block
Step B4.
For ease of understanding, here step B1-step B4 is illustrated: such as, for pending
Image obtains numbering after over-segmentation processes and is respectively 6 pending image blocks of A1-A6, through first
After secondary merging, it is assumed that the merging of the most numbered A1 and numbered A2 (image block after merging is designated as B1,
And put into optimization set), the merging of numbered A3 and numbered A4 (image block after merging is designated as B2,
And put into optimization set), numbering is respectively the image block of A5 with A6 and does not merges with adjacent pending image block
(all putting into optimization set), then, after the first round merges, have 4 pending images in optimization set
Block, numbering is respectively B1, B2, A5, A6.Repeat optimization set to merge, until optimization set
In the invariable number of pending image block (i.e. do not have aberration to be less than the adjacent pending image of default aberration
Block).After assuming to carry out first round merging, union operation also meets termination condition, the institute of the most pending image
Pending image block is had to include A1-A6 and newly-increased pending image block B1, B2.
Certainly, in one embodiment, in order to prevent excessively merging, cause pending image block excessive, this
In inventive embodiments, it is also possible in optimization set, the quantity of pending image block terminates less than after predetermined number
Union operation.That is, before performing step B4, it is thus necessary to determine that the number of pending image block in optimization set
Amount is less than predetermined number.So, merging the condition terminated is the quantity of pending image block in optimization set
Less than predetermined number.
Wherein, in one embodiment, the pending image block that obtains after above-mentioned over-segmentation being processed and new
After the pending image block increased is detected by category classifier, the object of the same area in pending image
May be detected repeatedly, that have probably due to the region difference of pending image block causes the accurate of testing result
Degree difference.So, in the embodiment of the present invention, in order to be accurately positioned out comprise in pending image every
The position of individual object, also includes the similar of the corresponding classification of pending image block in described testing result
Degree, described method also includes:
Step C1: according to each pending image block position in pending image, will be apart from less than presetting
The pending image block of distance is divided into one group.
Step C2: for the pending image block of least one set, chooses maximum pending of similarity in this group
Image block is as the representative image block of this group.
Wherein, step C2 is performed for the often group in the pending image block of least one set.
So, by the pending image block of close for position (i.e. distance is less than predeterminable range) is divided into one group,
And select the maximum pending image block of similarity as representative image block, i.e. by finding representative image block,
It is accurately positioned out in pending image the position of the object comprised.Certainly, when there being multiple image block to be grouped,
Multiple representative image block can be obtained, then each representative image block will comprise an object.
Obtaining after representative image block, step 103 is operated by the process of attributive classification device, can specifically perform be
By the attributive classification device corresponding with the classification determined, representative image block is detected, determine this representative graph
As the attribute that block is corresponding.So, the object of attributive classification device detection is only representative image block rather than owns
Pending image block, it is possible to reduce the quantity being carried out the image block detected by attributive classification device, at raising
Reason efficiency.
Wherein, in one embodiment, step 103 is performed (i.e. by the attribute corresponding with the classification determined
Grader, detects this pending image block, determines the attribute that this pending image block is corresponding) after,
All right: according to the attribute determined, choose from preset object information set and comprise the described attribute determined
Object information, using similar right as the object comprised in described pending image of the object information chosen
As.So, it is determined by analogical object, the homologue of the object comprised in pending image can be found out,
So that user obtains more valuable information by obtaining analogical object.
Wherein, in one embodiment, it is thus achieved that after analogical object, it is also possible to by described pending image
The analogical object of the object comprised is sent to described client and shows.So that user can be obtained by terminal
More valuable information.
Such as, in the embodiment of the present invention, when user plays video, can say in the video that user plays
Every two field picture is considered as pending image, and obtains description information and the phase of the object comprised in pending image
Show like object.Such as, the design sketch of display can be as it is shown on figure 3, show pending image in Fig. 3
In the numbering of each object that comprises, and the description information of each object and analogical object.User is permissible
By selecting one of them object as objects, by clicking on this objects, terminal is permissible
Description information and the information of analogical object, such as objects of launching this objects are clothes
Time, the information such as the brand of this clothes, price, and the purchase website of correspondence can be shown.So, Yong Hu
During viewing video, user can conveniently obtain the information of objects.
Such as, when user browses Web page image, it is possible to specify this image is pending image, and by client
It is sent to the device (such as server) with the embodiment of the present invention one function, by this device after analysis,
Description information and the analogical object of the object comprised in image are returned to client and show, and without user with
By means of the impression of oneself, use word removal search.Thus, the embodiment of the present invention can improve the efficiency of information processing.
Embodiment five
Based on identical inventive concept, the embodiment of the present invention also provides for a kind of information processor, such as Fig. 4 institute
Showing, described device includes:
Segmentation module 401, for carrying out image segmentation to pending image, it is thus achieved that at least one pending figure
As block；
Detection module 402, for for this pending image block, by category classifier to this process image
Block detects, it is thus achieved that testing result, and this testing result includes the classification that this pending image block is corresponding；
Attribute determination module 403, for by the attributive classification device corresponding with the classification determined, waits to locate to this
Reason image block detects, and determines the attribute that this pending image block is corresponding；
Information determination module 404 is described, for description information corresponding for the attribute determined being located as described waiting
The description information of the object comprised in reason image；
Sending module 405, comprises in the described pending image of acquisition ought receive client transmission
When the acquisition of the information of object is asked, described description information is sent to client and shows.
Wherein, in one embodiment, specifically include:
Cutting unit, for carrying out over-segmentation process to pending image, it is thus achieved that at least one destination object；
Image block determines unit, for for this destination object, determines and can comprise this mesh in pending image
The minimum preset pattern region of mark object is as a pending image block.
Wherein, in one embodiment, described device also includes:
Aberration determines module, determines that unit, for this destination object, determines in pending image for image block
Can comprise the minimum preset pattern region of this destination object as a pending image block after, for
The two pending image blocks that in pending image, position is adjacent, calculate the color of these adjacent two pending image blocks
Difference；
Merge module, if for aberration less than presetting aberration, then these adjacent two pending image blocks merged,
And the image block after merging is as newly-increased pending image block, and the image block after merging puts into optimization
In set；
Optimize module, if for aberration more than or equal to presetting aberration, then by these adjacent two pending image blocks
Put in optimization set；
Perform module, for for the pending image block in optimization set, trigger aberration and determine that module performs
For the two pending image blocks that position is adjacent in pending image, calculate this adjacent two pending image blocks
The step of aberration.
Wherein, in one embodiment, described device also includes:
Quantity determines module, for performing module for the pending image block in optimization set, triggering aberration
Determine module perform for the two pending image blocks that position is adjacent in pending image, calculate this adjacent two
Before the step of the aberration of pending image block, determine that in optimization set, the quantity of pending image block is less than pre-
If quantity.
Wherein, in one embodiment, described testing result also including, pending image block is corresponding
The similarity of classification, described device also includes:
Grouping module, for according to each pending image block position in pending image, will be apart from being less than
The pending image block of predeterminable range is divided into one group；
Representative image block determines unit, for for the pending image block of least one set, chooses in this group similar
The pending image block of degree maximum is as the representative image block of this group；
Attribute determination module, specifically for: by the attributive classification device corresponding with the classification determined, to representative
Image block detects, and determines the attribute that this representative image block is corresponding.
Wherein, in one embodiment, described device also includes:
Analogical object determines module, for attribute determination module by the attributive classification corresponding with the classification determined
Device, detects this pending image block, after determining the attribute that this pending image block is corresponding, according to
The attribute determined, chooses the object information comprising the described attribute determined from preset object information set,
Using the object information chosen as the analogical object of the object comprised in described pending image.
Wherein, in one embodiment, described device also includes:
Analogical object sending module, the analogical object of the object for comprising in described pending image sends
Show to described client.
Device embodiment described above is only schematically, wherein said illustrates as separating component
Unit can be or may not be physically separate, and the parts shown as unit can be or also
Can not be physical location, i.e. may be located at a place, or can also be distributed on multiple NE.
Some or all of module therein can be selected according to the actual needs to realize the mesh of the present embodiment scheme
's.Those of ordinary skill in the art, in the case of not paying performing creative labour, are i.e. appreciated that and implement.
Through the above description of the embodiments, those skilled in the art is it can be understood that arrive each enforcement
Mode can add the mode of required general hardware platform by software and realize, naturally it is also possible to pass through hardware.
Based on such understanding, the part that prior art is contributed by technique scheme the most in other words is permissible
Embodying with the form of software product, this computer software product can be stored in computer-readable storage medium
In matter, such as ROM/RAM, magnetic disc, CD etc., including some instructions with so that a computer equipment
(can be personal computer, server, or the network equipment etc.) performs each embodiment or embodiment
The method described in some part.
Last it is noted that above example is only in order to illustrate the technical scheme of the embodiment of the present invention, rather than
It is limited；Although the embodiment of the present invention being described in detail with reference to previous embodiment, this area
Those of ordinary skill is it is understood that the technical scheme described in foregoing embodiments still can be repaiied by it
Change, or wherein portion of techniques feature is carried out equivalent；And these amendments or replacement, do not make phase
The essence answering technical scheme departs from the spirit and scope of the embodiment of the present invention each embodiment technical scheme.
Claims (14)
1. an information processing method, it is characterised in that described method includes:
Pending image is carried out image segmentation, it is thus achieved that at least one pending image block；
For this pending image block, by category classifier, this process image block is detected, it is thus achieved that inspection
Surveying result, this testing result includes the classification that this pending image block is corresponding；And,
By the attributive classification device corresponding with the classification determined, this pending image block is detected, determines
The attribute that this pending image block is corresponding；And,
Using description information corresponding for the attribute determined as the description of the object comprised in described pending image
Information；
When the acquisition receiving the information obtaining the object comprised in described pending image that client sends
During request, described description information is sent to client and shows.
Method the most according to claim 1, it is characterised in that described pending image is carried out figure
As segmentation, it is thus achieved that at least one image block, specifically include:
Pending image is carried out over-segmentation process, it is thus achieved that at least one destination object；
For this destination object, determine the minimum preset pattern that can comprise this destination object in pending image
Region is as a pending image block.
Method the most according to claim 2, it is characterised in that for this destination object, determine and treat
Process in image and can comprise the minimum preset pattern region of this destination object as a pending image block
Afterwards, described method also includes:
For the two pending image blocks that position is adjacent in pending image, calculate this adjacent two pending figures
Aberration as block；
If aberration less than preset aberration, then by this adjacent two pending image blocks merge, and will merge after figure
As block is as newly-increased pending image block, and the image block after merging is put in optimization set；
If aberration is more than or equal to presetting aberration, then these adjacent two pending image blocks are put into optimization set
In；
For the pending image block in optimization set, return execution adjacent for position in pending image
Two pending image blocks, calculate the step of the aberration of these adjacent two pending image blocks.
Method the most according to claim 3, it is characterised in that pending in optimization set
Image block, returns and performs for the two pending image blocks that position is adjacent in pending image, calculate this phase
Before the step of the aberration of adjacent two pending image blocks, described method also includes:
Determine that in optimization set, the quantity of pending image block is less than predetermined number.
5. according to described method arbitrary in claim 2-4, it is characterised in that in described testing result
Also including the similarity of the corresponding classification of pending image block, described method also includes:
According to each pending image block position in pending image, by distance less than predeterminable range wait locate
Reason image block is divided into one group；
For the pending image block of least one set, choose the pending image block conduct that in this group, similarity is maximum
The representative image block of this group；
Described by the attributive classification device corresponding with the classification determined, this pending image block is detected,
Determine the attribute that this pending image block is corresponding, specifically include:
By the attributive classification device corresponding with the classification determined, representative image block is detected, determine this generation
The attribute that table image block is corresponding.
Method the most according to claim 1, it is characterised in that by corresponding with the classification determined
Attributive classification device, detects this pending image block, determine attribute that this pending image block is corresponding it
After, described method also includes:
According to the attribute determined, from preset object information set, choose the thing comprising the described attribute determined
Body information, using the object information chosen as the analogical object of the object comprised in described pending image.
Method the most according to claim 6, it is characterised in that described method also includes:
The analogical object of the object comprised in described pending image is sent to described client show.
8. an information processor, it is characterised in that described device includes:
Segmentation module, for carrying out image segmentation to pending image, it is thus achieved that at least one pending image block；
Detection module, for for this pending image block, is entered this process image block by category classifier
Row detection, it is thus achieved that testing result, this testing result includes the classification that this pending image block is corresponding；
Attribute determination module, for by the attributive classification device corresponding with the classification determined, to this pending figure
As block detects, determine the attribute that this pending image block is corresponding；
Information determination module is described, for by description information corresponding for the attribute that determine as described pending figure
The description information of the object comprised in Xiang；
Sending module, for when receiving the object comprised in the described pending image of acquisition that client sends
Information acquisition request time, described description information is sent to client and shows.
Device the most according to claim 8, it is characterised in that described segmentation module, specifically includes:
Cutting unit, for carrying out over-segmentation process to pending image, it is thus achieved that at least one destination object；
Image block determines unit, for for this destination object, determines and can comprise this mesh in pending image
The minimum preset pattern region of mark object is as a pending image block.
Device the most according to claim 9, it is characterised in that described device also includes:
Aberration determines module, determines that unit, for this destination object, determines in pending image for image block
Can comprise the minimum preset pattern region of this destination object as a pending image block after, for
The two pending image blocks that in pending image, position is adjacent, calculate the color of these adjacent two pending image blocks
Difference；
Merge module, if for aberration less than presetting aberration, then these adjacent two pending image blocks merged,
And the image block after merging is as newly-increased pending image block, and the image block after merging puts into optimization
In set；
Optimize module, if for aberration more than or equal to presetting aberration, then by these adjacent two pending image blocks
Put in optimization set；
Perform module, for for the pending image block in optimization set, trigger aberration and determine that module performs
For the two pending image blocks that position is adjacent in pending image, calculate this adjacent two pending image blocks
The step of aberration.
11. devices according to claim 10, it is characterised in that described device also includes:
Quantity determines module, for performing module for the pending image block in optimization set, triggering aberration
Determine module perform for the two pending image blocks that position is adjacent in pending image, calculate this adjacent two
Before the step of the aberration of pending image block, determine that in optimization set, the quantity of pending image block is less than pre-
If quantity.
12. according to described device arbitrary in claim 9-11, it is characterised in that in described testing result
Also including the similarity of the corresponding classification of pending image block, described device also includes:
Grouping module, for according to each pending image block position in pending image, will be apart from being less than
The pending image block of predeterminable range is divided into one group；
Representative image block determines unit, for for the pending image block of least one set, chooses in this group similar
The pending image block of degree maximum is as the representative image block of this group；
Attribute determination module, specifically for: by the attributive classification device corresponding with the classification determined, to representative
Image block detects, and determines the attribute that this representative image block is corresponding.
13. devices according to claim 8, it is characterised in that described device also includes:
Analogical object determines module, for attribute determination module by the attributive classification corresponding with the classification determined
Device, detects this pending image block, after determining the attribute that this pending image block is corresponding, according to
The attribute determined, chooses the object information comprising the described attribute determined from preset object information set,
Using the object information chosen as the analogical object of the object comprised in described pending image.
14. devices according to claim 13, it is characterised in that described device also includes:
Analogical object sending module, the analogical object of the object for comprising in described pending image sends
Show to described client.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610281956.2A CN105975922A (en) | 2016-04-29 | 2016-04-29 | Information processing method and information processing device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610281956.2A CN105975922A (en) | 2016-04-29 | 2016-04-29 | Information processing method and information processing device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105975922A true CN105975922A (en) | 2016-09-28 |
Family
ID=56993513
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610281956.2A Pending CN105975922A (en) | 2016-04-29 | 2016-04-29 | Information processing method and information processing device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105975922A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108122238A (en) * | 2018-01-30 | 2018-06-05 | 百度在线网络技术（北京）有限公司 | Image processing method, device, equipment and computer readable storage medium |
| CN110598705A (en) * | 2019-09-27 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Semantic annotation method and device for image |
| CN111091158A (en) * | 2019-12-25 | 2020-05-01 | 科大讯飞股份有限公司 | Method, device and equipment for classifying image quality of auxiliary images |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521565A (en) * | 2011-11-23 | 2012-06-27 | 浙江晨鹰科技有限公司 | Garment identification method and system for low-resolution video |
| CN103377376A (en) * | 2012-04-13 | 2013-10-30 | 阿里巴巴集团控股有限公司 | Method and system for image classification, and method and system for image retrieval |
| CN103530405A (en) * | 2013-10-23 | 2014-01-22 | 天津大学 | Image retrieval method based on layered structure |
| CN104504055A (en) * | 2014-12-19 | 2015-04-08 | 常州飞寻视讯信息科技有限公司 | Commodity similarity calculation method and commodity recommending system based on image similarity |
| CN105260747A (en) * | 2015-09-30 | 2016-01-20 | 广东工业大学 | Clothing identification method based on clothes concurrent information and multitask learning |
| CN105389588A (en) * | 2015-11-04 | 2016-03-09 | 上海交通大学 | Multi-semantic-codebook-based image feature representation method |
| CN105426455A (en) * | 2015-11-12 | 2016-03-23 | 中国科学院重庆绿色智能技术研究院 | Method and device for carrying out classified management on clothes on the basis of picture processing |
| CN105469087A (en) * | 2015-07-13 | 2016-04-06 | 百度在线网络技术（北京）有限公司 | Method for identifying clothes image, and labeling method and device of clothes image |
| CN105475529A (en) * | 2016-01-05 | 2016-04-13 | 黄阿福 | Cooked and loose Pu'er tea manufacturing technology |
-
2016
- 2016-04-29 CN CN201610281956.2A patent/CN105975922A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521565A (en) * | 2011-11-23 | 2012-06-27 | 浙江晨鹰科技有限公司 | Garment identification method and system for low-resolution video |
| CN103377376A (en) * | 2012-04-13 | 2013-10-30 | 阿里巴巴集团控股有限公司 | Method and system for image classification, and method and system for image retrieval |
| CN103530405A (en) * | 2013-10-23 | 2014-01-22 | 天津大学 | Image retrieval method based on layered structure |
| CN104504055A (en) * | 2014-12-19 | 2015-04-08 | 常州飞寻视讯信息科技有限公司 | Commodity similarity calculation method and commodity recommending system based on image similarity |
| CN105469087A (en) * | 2015-07-13 | 2016-04-06 | 百度在线网络技术（北京）有限公司 | Method for identifying clothes image, and labeling method and device of clothes image |
| CN105260747A (en) * | 2015-09-30 | 2016-01-20 | 广东工业大学 | Clothing identification method based on clothes concurrent information and multitask learning |
| CN105389588A (en) * | 2015-11-04 | 2016-03-09 | 上海交通大学 | Multi-semantic-codebook-based image feature representation method |
| CN105426455A (en) * | 2015-11-12 | 2016-03-23 | 中国科学院重庆绿色智能技术研究院 | Method and device for carrying out classified management on clothes on the basis of picture processing |
| CN105475529A (en) * | 2016-01-05 | 2016-04-13 | 黄阿福 | Cooked and loose Pu'er tea manufacturing technology |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108122238A (en) * | 2018-01-30 | 2018-06-05 | 百度在线网络技术（北京）有限公司 | Image processing method, device, equipment and computer readable storage medium |
| CN110598705A (en) * | 2019-09-27 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Semantic annotation method and device for image |
| CN111091158A (en) * | 2019-12-25 | 2020-05-01 | 科大讯飞股份有限公司 | Method, device and equipment for classifying image quality of auxiliary images |
| CN111091158B (en) * | 2019-12-25 | 2024-04-30 | 科大讯飞股份有限公司 | Classification method, device and equipment for image quality of teaching auxiliary image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105956059A (en) | Emotion recognition-based information recommendation method and apparatus | |
| CN109670591B (en) | Neural network training method and image matching method and device | |
| CN104239909B (en) | A kind of recognition methods of image and device | |
| CN102915372B (en) | Image search method, Apparatus and system | |
| CN104899586B (en) | Method and device is identified to the word content for including in image | |
| CN106960195A (en) | A kind of people counting method and device based on deep learning | |
| CN106227827B (en) | Image of clothing foreground color feature extracting method and costume retrieval method and system | |
| AU2014321165A1 (en) | Image searching method and apparatus | |
| CN105493078B (en) | Colored sketches picture search | |
| CN106649610A (en) | Image labeling method and apparatus | |
| CN108205676B (en) | The method and apparatus for extracting pictograph region | |
| KR20160120674A (en) | Clothes recommendation system using gpu | |
| CN111144215B (en) | Image processing method, device, electronic equipment and storage medium | |
| CN108629319B (en) | Image detection method and system | |
| JPH05503798A (en) | Touching character identification and segmentation method based on contour characteristics | |
| CN107330360A (en) | A kind of pedestrian's clothing colour recognition, pedestrian retrieval method and device | |
| CN111212303A (en) | Video recommendation method, server and computer-readable storage medium | |
| CN106372111A (en) | Local feature point screening method and system | |
| CN108537286A (en) | A kind of accurate recognition methods of complex target based on key area detection | |
| CN105975922A (en) | Information processing method and information processing device | |
| US20180276460A1 (en) | Information processing apparatus | |
| CN108492160A (en) | Information recommendation method and device | |
| CN107301408A (en) | Human body mask extracting method and device | |
| KR20220039578A (en) | Method for providing clothing recommendation information based on user-selected clothing, and server and program using the same | |
| CN111222530A (en) | Fine-grained image classification method, system, device and storage medium |
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: 20160928 |
|
| WD01 | Invention patent application deemed withdrawn after publication |