CN108346129A - Generating has the method for the picture mosaic segment for obscuring segment - Google Patents
Generating has the method for the picture mosaic segment for obscuring segment Download PDFInfo
- Publication number
- CN108346129A CN108346129A CN201810199413.5A CN201810199413A CN108346129A CN 108346129 A CN108346129 A CN 108346129A CN 201810199413 A CN201810199413 A CN 201810199413A CN 108346129 A CN108346129 A CN 108346129A
- Authority
- CN
- China
- Prior art keywords
- segment
- picture
- style
- synthesising
- obscuring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000006243 chemical reaction Methods 0.000 claims abstract description 6
- 230000004438 eyesight Effects 0.000 claims description 7
- 230000000007 visual effect Effects 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 238000012546 transfer Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 2
- 230000004304 visual acuity Effects 0.000 abstract description 3
- 238000013507 mapping Methods 0.000 description 17
- 238000010586 diagram Methods 0.000 description 7
- 239000002131 composite material Substances 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 208000003464 asthenopia Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000010422 painting Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Classifications
-
- G06T3/04—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
Abstract
The present invention provides a kind of method for generating and having the picture mosaic segment for obscuring segment, including：For each in multiple original images, picture style conversion is carried out to it using one in multiple style templates, obtains corresponding synthesising picture；Each in obtained multiple synthesising pictures is divided into multiple segments with predetermined size；And for the segment with similar style, clustered based on content similarity, and for all figures each segment in the block under the same category, other segments under the category are obscured into segment as the segment.The present invention will obscure item and introduce picture mosaic segment, to promote observation, resolving power and the imagination of user to a greater degree in the limited screen space of electronic equipment.
Description
Technical field
The present invention relates to information service and image processing field more particularly to the generation techniques of picture mosaic segment.
Background technology
Picture mosaic is a kind of intelligence tool being commonly used in training and exploitation children observation, resolving power and imagination,
Multiple segments by that can be combined into a width full picture form.With the development of electronic technology, children can be more easily in hand
Intellectual development is carried out by picture mosaic application on the electronic equipments such as machine, tablet computer.In entity picture mosaic field, expand picture mosaic picture
Picture mosaic picture is subdivided into the conventional means that more thumbnails are raising picture mosaic difficulty by size.However, due to intelligent hand
Machine, the screen size of tablet computer are limited, thus large-sized picture mosaic picture show it is very inconvenient.In addition, if segment
It is too small, it is unfavorable for operating and easy tos produce visual fatigue.
Invention content
To overcome above-mentioned defect existing in the prior art, the present invention to provide following technical scheme：
In one aspect, a kind of method for generating and having the picture mosaic segment for obscuring segment is provided, including：
Step 1) carries out picture for each in multiple original images, using one in multiple style templates to it
Style is converted, and corresponding synthesising picture is obtained；
Each in obtained multiple synthesising pictures is divided into multiple segments with predetermined size by step 2),
In, the segment in the synthesising picture obtained using identical style template is regarded as with similar style；
Step 3) clusters based on content similarity to obtain the class belonging to segment the segment with similar style
Not, and for all figures each segment in the block under the same category, using other segments under the category as the segment
Obscure segment.
In the above method, in step 3), the segment with similar style is clustered based on content similarity can
To include：For each segment extraction local feature region with similar style and the local feature region extracted is gathered
The center of each cluster is corresponded to a vision word, the visual dictionary of the segment is characterized with structure by class；And for regarding
Feel that all segments with similar style of dictionary characterization are clustered, obtains the classification belonging to segment.
In the above method, in step 1), carrying out the conversion of picture style to original image using style template may include：
Using the picture style transfer algorithm based on convolutional neural networks, original image is converted into having with the style template utilized
The synthesising picture of similar style.
In the above method, can also include before step 1)：Original image is pre-processed, so that its length and width
Degree meets scheduled size.
The above method can also include：Record each synthesising picture and between the segment being partitioned into the synthesising picture
Correspondence；And record classification belonging to each segment.
On the other hand, a kind of method for providing a user and including the picture mosaic segment for obscuring segment is provided, including：
Multiple synthesising pictures are provided a user to be selected by user, the multiple synthesising picture is to state method life in use
At the multiple synthesising pictures obtained during with the picture mosaic segment for obscuring segment；
According to the synthesising picture that user selects, from obscuring the picture mosaic of segment with being found in segment from the synthesising picture
All segments being partitioned into are as first group of segment；
From it is described have obscure the picture mosaic of segment and obscure figure with find each segment in first group of segment in segment
Block is as second group of segment；
First group of segment and second group of segment are supplied to user together.
Compared with the prior art, the advantages of the present invention are as follows：
1. without expanding the size of picture mosaic picture or reducing the size of segment, answered suitable for picture mosaic on an electronic device
With；
2. being provided while providing a user picture mosaic segment and obscuring item, it is relatively beneficial to training and the sight of development of user
Examine power, resolving power and imagination.
Description of the drawings
Embodiments of the present invention is further illustrated referring to the drawings, wherein：
Fig. 1 is the method flow diagram according to an embodiment of the invention for generating and having the picture mosaic segment for obscuring segment；
Fig. 2 is the side that the segment consistent to style based on content similarity according to an embodiment of the invention is classified
Method flow chart；
Fig. 3 is the method stream according to an embodiment of the invention for providing a user and including the picture mosaic segment for obscuring segment
Cheng Tu；
Fig. 4 is according to an embodiment of the invention by generating based on the method for picture mosaic segment for having and obscuring segment
The schematic diagram of calculation machine device；
Fig. 5 (a) is an example of original image；
Fig. 5 (b) is an example of style template；
Fig. 5 (c) is to carry out picture style conversion to the original image in Fig. 5 (a) using style template shown in Fig. 5 (b)
The obtained synthesising picture consistent with Fig. 5 (b) styles.
Specific implementation mode
In order to make the purpose of the present invention, technical solution and advantage be more clearly understood, pass through below in conjunction with attached drawing specific real
Applying example, the present invention is described in more detail.It should be appreciated that described herein, specific examples are only used to explain the present invention, and
It is not used in the restriction present invention.
It should be noted that some illustrative methods are depicted as flow chart.It is executed although operation is expressed as sequence by flow chart,
But it is understood that many operations can be parallel while or synchronously being executed.Furthermore it is possible to rearrange the sequence of operation.
Processing can be terminated operating the when of completing, but can also be had and be not included in the other step in figure or in embodiment.
Fig. 4 shows a kind of signal for generating the computer installation of the method with the picture mosaic segment for obscuring segment
Figure a comprising processor unit and used data storage mechanism during executing this method, the latter include：
Non-standard original graph valut, standard original graph valut, style template library, composite diagram valut, synthesising picture segment library, style template
Record sheet, synthesising picture mapping table, segment mapping table and information scratchpad area (SPA).
Wherein, non-standard original graph valut does not meet the original image of predetermined specifications for memory length and width；Standard
Original graph valut is for storing the original image that image length and width meet predetermined specifications；Style picture library is for storing multiple style moulds
Plate；Composite diagram valut is used to store carries out the synthesising picture that style is converted to using style template to original image；Composite diagram
Piece segment library is used to store the segment obtained after synthesising picture segmentation；And information scratchpad area (SPA) is joined for storage system configuration
The ephemeral data generated in number and processing procedure.In addition, style template record table is used to record the mark of style template (for only
One identifies the style template, and filename can be used for convenience) and its storage location, as shown in Table 1 below；Synthesising picture
(for the unique mark synthesising picture, file can be used in the mark that mapping table is used to store synthesising picture for convenience
Name) and its storage location and style template and synthesising picture correspondence, as shown in Table 2 below；Segment correspondence
Table is used to store the mark (for the unique mark segment, filename can be used for convenience) of segment, classification logotype and deposits
Storage space is set and the correspondence of segment and synthesising picture, as shown in Table 3 below.Table 1：Style template record table
| Style template identification | Style template storage location |
Table 2：Synthesising picture mapping table
| Synthesising picture identifies | Style template identification | Synthesising picture storage location |
Table 3：Segment mapping table
| Synthesising picture identifies | Segment identifies | Classification logotype | Segment storage location |
Fig. 1 shows the method according to an embodiment of the invention for generating and having the picture mosaic segment for obscuring segment.Generally
For including, this method includes：Step S11, for each original image, using one in multiple and different style templates to it
Picture style conversion is carried out, to obtain corresponding synthesising picture；Obtained each synthesising picture is divided into tool by step S12
There are multiple segments of predetermined size；And step S13 gathers it based on content similarity the consistent segment of style
Class, and all segments under the same category are regarded as and obscure segment each other.Hereafter this method will be unfolded to be described in detail：
In step s 11, each original image (such as Fig. 5 of predetermined specifications is met for size in standard original graph valut
(a) shown in), a style is selected from multiple style templates (or style picture, template picture etc.) in style template library
Template (shown in such as Fig. 5 (b)) converts (Image Style to the original image using the picture style based on convolutional neural networks
Transfer or picture Style Transfer) algorithm, with obtain it is corresponding with the original image and with the style mould that is utilized
The consistent synthesising picture of plate style, as shown in Fig. 5 (c).Wherein, size meets the original images of predetermined specifications and is stored in such as Fig. 4
Shown in standard original graph valut, and style template is stored in style template library as shown in Figure 4.For above-mentioned original
For picture, as known to the skilled person, feature includes style and content.The style of picture is different from the interior of picture
Hold, the latter refers to the things reflected in picture, and what the style of picture was often referred to that the picture shows on the whole has
Representative uniqueness looks, such as painting style, sketch style, cartoon style etc..With the development of depth learning technology, volume
On the one hand product neural network can extract the high-level characteristic of picture so that the style and content of picture detach, another party
Face can use the style extracted expression to render the content of other pictures.Therefore, above-mentioned to utilize style template to original
Beginning picture applies the picture style transfer algorithm based on convolutional neural networks, refers to utilizing style mould by convolutional neural networks
The style of plate is expressed to render the content of original image, thus the style in the original image after rendering is carried with from style template
The style of taking-up is consistent.
After carrying out picture style to original image using style template and being converted to synthesising picture, also by the synthesis
Picture is stored in composite diagram valut, also, also by the correspondence between the style template utilized and the synthesising picture
It is stored in synthesising picture mapping table as shown in Table 2, in order to be used in subsequent step.
It should be noted that the size of original image needs to meet predetermined specifications, i.e., obtained from original image (that is, through
Picture style is converted to) corresponding synthesising picture can just be divided into multiple segments with predetermined size.For example, it is assumed that
The predetermined size of segment is 2cm × 2cm, can be from corresponding synthesising picture if the size of original image is 8cm × 6cm
In be just partitioned into 12 segments；If the size of original image is 8cm × 8cm, it can just be partitioned into 16 segments；So
And if the size of original image is 8cm × 7cm, it can not be partitioned into the segment of all 2cm × 2cm of size, in this feelings
It under condition, needs first to cut original image, such as be cut to the picture of 8cm × 6cm, so as to just be partitioned into 12
Segment.
In step s 12, each synthesising picture is divided into multiple segments with predetermined size, for example, by 8cm ×
The synthesising picture of 6cm is divided into the segment of 12 2cm × 2cm, to obtain the picture mosaic suitable for electronic equipment with segment (about
How to determine that the segment of obscuring of each segment will hereinafter be described).As described above, style template and being obtained using it
Synthesising picture style is consistent, therefore herein, regards all synthesising pictures obtained using identical style template as style
Consistent.In addition, it is consistent that the segment divided from these synthesising pictures is also considered as style, that is, identical wind will be utilized
All segments in all synthesising pictures that grid template obtains regard the consistent segment of style as.
After being split to obtain corresponding segment to synthesising picture, which can be stored in synthesising picture segment library
In, and the correspondence between the synthesising picture and corresponding segment can be recorded in segment mapping table, in order to
It is used in subsequent step.For example, can be obtained with segmentation using segment mapping table as shown in table 3 to record synthesising picture
Segment between correspondence.
In step s 13, for the consistent segment of style, it is clustered based on content similarity, and for same
Other segments under the category are obscured segment by all figure each segments in the block as the segment under one classification.Wherein, may be used
Own using what the same style template obtained to be found according to above-mentioned synthesising picture mapping table and segment mapping table
Segment, such as definition above, these segment styles are consistent, (or these segments have similar style, that is, from a style mould
All segments that plate obtains belong to similar style).For the segment under every a kind of style, according to one embodiment of present invention,
The cluster based on content similarity is carried out using method shown in Fig. 2, this method includes：
In the step s 21, for each segment with such style, with SIFT (Scale-invariant feature
Transform, scale invariant feature conversion) or SURF (Speeded Up Robust Features accelerate robust feature) operator
Local feature region is extracted, the feature vector for each local feature region same dimension extracted is indicated.
In step S22, utilizing k-means algorithms for the local feature region of each segment with such style, (K is equal
Value clusters) it is clustered, obtained each cluster centre is corresponded into a vision word, regarding for the segment is characterized to construct
Feel dictionary.Wherein, each vision word is a digital vectors, and dimension for example can be 128 dimensions.The dimension of visual dictionary is
Every dimension of the quantity (for example, 800) of vision word, visual dictionary corresponds to a vision word.Characterizing some segment
Visual dictionary in, the numerical value per dimension can be corresponding vision word in figure occurrence number in the block, or make
It is obtained with TF-IDF (Term Frequency-Inverse Document Frequency, word frequency-inverse file frequency) method
Numerical value.
In step S23, by all segments with such style characterized with visual dictionary using k-means algorithms into
Row cluster, obtains the classification belonging to each segment, records the classification belonging to each segment.As set forth above, it is possible to by each segment
Affiliated classification is recorded in segment mapping table (that is, table 3 above).
It should be noted that in step s 13, concurrently step S21- can be executed to the segment under every a kind of style
S23, if being all to identify its classification since number 1 for the segment under every a kind of style, segment mapping table (on
The table 3 of text) in, the classification logotype for not having the segment (that is, the segment obtained using different-style template) of similar style may
Conflict is had, then may subsequently also need to synthesising picture mapping table (table 2 above) and coordinate together to determine segment with table 3
Obscure segment.
After obtaining the classification of segment, the segment of the same category is regarded as and obscures segment each other (i.e., it is determined that segment
Obscure segment, generated as a result, with the picture mosaic segment for obscuring segment) because these segments are of the same size, unanimously
Style and similar content so that the more difficult resolution of human eye.It, can will be identical when submitting picture mosaic segment to user
Other segments conduct under classification obscures item and is supplied to user together, to improve the difficulty of picture mosaic.It should be noted that right
After segment is clustered, some classifications may be there is only a segment, then these segments do not obscure item for what is obscured.
The foregoing describe one embodiment that generation as shown in Figure 1 has the method for the picture mosaic segment for obscuring segment,
Can also include pre-treatment step before step S11 according to another embodiment of this method.That is, to non-standard original graph
Size does not meet the original images of predetermined specifications and is handled in valut, so that its length and width meets predetermined specifications (that is, making
Multiple segments with predetermined size can just be divided by obtaining the corresponding synthesising picture of the original image), and will be through handling
To size meet the original images of predetermined specifications and be stored in standard original graph valut as shown in Figure 4.
In addition, it should also be understood by those skilled in the art that although some specific algorithms or method is used above to describe to give birth to
At the method with the picture mosaic segment for obscuring segment, but these algorithms or method are not restricted herein, for example, can utilize
Histogram Matching calculates the similarity of segment content, segment can also classify using density clustering etc..
Fig. 3 shows the side according to an embodiment of the invention for providing a user and including the picture mosaic segment for obscuring segment
Method includes the following steps：
In step S31, the synthesising picture that can be used for picture mosaic is provided a user, to be selected by users.As described above,
The synthesising picture be in use text described in method generate have obscure segment picture mosaic segment when, to original image carry out
The synthesising picture that picture style is converted to can find these synthesising pictures from composite diagram valut.
In step s 32, according to user select synthesising picture, from according to method as discussed above generate have obscure
The picture mosaic of segment, which is used, finds all segments for being partitioned into from the synthesising picture as first group of segment in segment.For example, can root
The synthesising picture mark submitted according to user, from finding all figures being partitioned into from the synthesising picture in segment mapping table
Block, and find the classification belonging to each segment.
In step S33, obscures the picture mosaic of segment from having for generation and each schemed with being found in segment in first group of segment
Block obscures segment as second group of segment.Specifically, above-mentioned synthesising picture can be found from synthesising picture mapping table
Corresponding style template identification, and according to the style template identification, found in synthesising picture mapping table and utilize the wind
All synthesising pictures mark that grid template obtains；Then, it finds in segment mapping table and is obtained using the style template
Segment corresponding to all synthesising pictures, using these segments as second group of alternative segment；Finally, according to segment correspondence
Table selects classification segment corresponding with the classification of the first group picture segment in the block, as second from second group of alternative segment
Group segment.
In step S34, first group of segment and second group of segment are supplied to user together.
The above method can pass through hardware, software, firmware, middleware, pseudocode, hardware description language or their times
Meaning combines to realize.When being implemented with software, firmware, middleware or pseudocode, for executing the program code or code of task
Segmentation can be stored in computer-readable medium, and such as storage medium, processor can execute the task.
It should be understood that the exemplary embodiment of software realization usually carried out on some form of program recorded medium coding or
Person realizes on some type of transmission medium.Program recorded medium can be arbitrary non-transitory storage media, such as disk
(for example, floppy disk or hard disk) or CD (for example, compact disk read-only memory or " CD ROM "), and can be it is read-only or
Random access.Similarly, transmission medium can be twisted-pair feeder, coaxial cable, optical fiber or known in the art some are other
Applicable transmission medium.
Although the present invention has been described by means of preferred embodiments, the present invention is not limited to described here
Embodiment, further include made various changes and variation without departing from the present invention.
Claims (8)
1. a kind of method that generation has the picture mosaic segment for obscuring segment, including：
Step 1) carries out picture style for each in multiple original images, using one in multiple style templates to it
Conversion, obtains corresponding synthesising picture；
Each in obtained multiple synthesising pictures is divided into multiple segments with predetermined size by step 2), wherein
Segment in the synthesising picture obtained using identical style template is regarded as with similar style；
Step 3) clusters based on content similarity to obtain the classification belonging to segment the segment with similar style,
And for all figures each segment in the block under the same category, using other segments obscuring as the segment under the category
Segment.
2. according to the method described in claim 1, in step 3), content similarity is based on for the segment with similar style
Carrying out cluster includes：
For each segment extraction local feature region with similar style and the local feature region extracted is clustered, it will
The center each clustered corresponds to a vision word, and the visual dictionary of the segment is characterized with structure；
All segments with similar style characterized with visual dictionary are clustered, obtain the classification belonging to segment.
3. method according to claim 1 or 2, in step 1), picture wind is carried out to original image using style template
Lattice are converted：
Using the picture style transfer algorithm based on convolutional neural networks, the style template that original image is converted into and is utilized
Synthesising picture with similar style.
4. method according to claim 1 or 2, wherein further include before step 1)：
Original image is pre-processed, so that its length and width meets scheduled size.
5. method according to claim 1 or 2, further including：
Record each synthesising picture and from the correspondence between the segment being partitioned into the synthesising picture；And
Record the classification belonging to each segment.
6. a kind of computer equipment, including processor and memory, the memory, which is stored with, to be executed by the processor
Instruction, make the computer equipment realize as any in claim 1-5 when described instruction is executed by the processor
Method described in.
7. it is a kind of provide a user comprising obscure segment picture mosaic segment method, including：
Multiple synthesising pictures are provided a user to be selected by user, the multiple synthesising picture is using such as claim 1-5
Any one of described in method generate have obscure the picture mosaic segment of segment during obtained multiple synthesising pictures；
According to the synthesising picture that user selects, from being divided from the synthesising picture with being found in segment with obscuring the picture mosaic of segment
All segments gone out are as first group of segment；
From described there is the picture mosaic for obscuring segment to be made with the segment of obscuring for finding each segment in first group of segment in segment
For second group of segment；
First group of segment and second group of segment are supplied to user together.
8. a kind of computer equipment, including processor and memory, the memory, which is stored with, to be executed by the processor
Instruction, make the computer equipment realize side as claimed in claim 7 when described instruction is executed by the processor
Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810199413.5A CN108346129B (en) | 2018-03-12 | 2018-03-12 | Method for generating puzzle blocks with confusing blocks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810199413.5A CN108346129B (en) | 2018-03-12 | 2018-03-12 | Method for generating puzzle blocks with confusing blocks |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108346129A true CN108346129A (en) | 2018-07-31 |
| CN108346129B CN108346129B (en) | 2020-07-31 |
Family
ID=62958103
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810199413.5A Active CN108346129B (en) | 2018-03-12 | 2018-03-12 | Method for generating puzzle blocks with confusing blocks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108346129B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1710596A (en) * | 2005-06-16 | 2005-12-21 | 上海交通大学 | Computer picture-arrangement automatic identifying method based on machine vision |
| CN201871218U (en) * | 2010-11-11 | 2011-06-22 | 田贵州 | Jigsaw |
| CN104915673A (en) * | 2014-03-11 | 2015-09-16 | 株式会社理光 | Object classification method and system based on bag of visual word model |
| CN105323065A (en) * | 2014-07-21 | 2016-02-10 | 腾讯科技（深圳）有限公司 | Safety verification method and device |
| CN106023081A (en) * | 2016-05-21 | 2016-10-12 | 广东邦宝益智玩具股份有限公司 | Mosaic processing method of 2D picture |
| CN107133920A (en) * | 2017-06-13 | 2017-09-05 | 华侨大学 | A kind of automatic generation method of the mosaic of view-based access control model feature |
| US20170259165A1 (en) * | 2016-03-08 | 2017-09-14 | James L. Prentice | Process for creating multiple-in-one jigsaw puzzles and for creating artwork from single images |
-
2018
- 2018-03-12 CN CN201810199413.5A patent/CN108346129B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1710596A (en) * | 2005-06-16 | 2005-12-21 | 上海交通大学 | Computer picture-arrangement automatic identifying method based on machine vision |
| CN201871218U (en) * | 2010-11-11 | 2011-06-22 | 田贵州 | Jigsaw |
| CN104915673A (en) * | 2014-03-11 | 2015-09-16 | 株式会社理光 | Object classification method and system based on bag of visual word model |
| CN105323065A (en) * | 2014-07-21 | 2016-02-10 | 腾讯科技（深圳）有限公司 | Safety verification method and device |
| US20170259165A1 (en) * | 2016-03-08 | 2017-09-14 | James L. Prentice | Process for creating multiple-in-one jigsaw puzzles and for creating artwork from single images |
| CN106023081A (en) * | 2016-05-21 | 2016-10-12 | 广东邦宝益智玩具股份有限公司 | Mosaic processing method of 2D picture |
| CN107133920A (en) * | 2017-06-13 | 2017-09-05 | 华侨大学 | A kind of automatic generation method of the mosaic of view-based access control model feature |
Non-Patent Citations (1)
| Title |
|---|
| 刘宁: "基于CBIR技术的计算机拼图教育软件平台设计", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108346129B (en) | 2020-07-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Uijlings et al. | Video classification with densely extracted hog/hof/mbh features: an evaluation of the accuracy/computational efficiency trade-off | |
| Wu et al. | Harvesting discriminative meta objects with deep CNN features for scene classification | |
| Naikal et al. | Informative feature selection for object recognition via sparse PCA | |
| EP3084682B1 (en) | System and method for identifying faces in unconstrained media | |
| RU2668717C1 (en) | Generation of marking of document images for training sample | |
| US9036905B2 (en) | Training classifiers for deblurring images | |
| US8699789B2 (en) | Document classification using multiple views | |
| Tasse et al. | Cluster-based point set saliency | |
| JP7270013B2 (en) | Information processing method, information processing device, electronic device and storage medium | |
| Ashraf et al. | Content-based Image Retrieval by Exploring Bandletized Regions through Support Vector Machines. | |
| US20180046721A1 (en) | Systems and Methods for Automatic Customization of Content Filtering | |
| JP4926266B2 (en) | Learning data creation device, learning data creation method and program | |
| CN108345700B (en) | Article representative picture selection method and device and computer equipment | |
| Liao et al. | An oil painters recognition method based on cluster multiple kernel learning algorithm | |
| Annum et al. | Saliency detection using contrast enhancement and texture smoothing operations | |
| Xu et al. | Automated classification for HEp-2 cells based on linear local distance coding framework | |
| CN108241861A (en) | A kind of data visualization method and equipment | |
| Battiato et al. | Scene categorization using bag of textons on spatial hierarchy | |
| Ye et al. | Saliency detection via similar image retrieval | |
| Saeed et al. | Combining ear-based traditional and soft biometrics for unconstrained ear recognition | |
| Lazebnik et al. | An empirical bayes approach to contextual region classification | |
| CN108346129A (en) | Generating has the method for the picture mosaic segment for obscuring segment | |
| Singh et al. | A review of different content based image retrieval techniques | |
| Yanagisawa et al. | Manga character clustering with DBSCAN using fine-tuned CNN model | |
| KR102444172B1 (en) | Method and System for Intelligent Mining of Digital Image Big-Data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |