CN106898035A - A kind of dress ornament sample set creation method and device - Google Patents
A kind of dress ornament sample set creation method and device Download PDFInfo
- Publication number
- CN106898035A CN106898035A CN201710038546.XA CN201710038546A CN106898035A CN 106898035 A CN106898035 A CN 106898035A CN 201710038546 A CN201710038546 A CN 201710038546A CN 106898035 A CN106898035 A CN 106898035A
- Authority
- CN
- China
- Prior art keywords
- dress ornament
- color
- sample set
- colour atla
- collection
- 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
- 238000000034 method Methods 0.000 title claims abstract description 26
- 241000532370 Atla Species 0.000 claims abstract description 85
- 239000003086 colorant Substances 0.000 claims abstract description 48
- 238000004040 coloring Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims description 14
- 238000005070 sampling Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 6
- 230000007547 defect Effects 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004043 dyeing Methods 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000037303 wrinkles Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/40—Filling a planar surface by adding surface attributes, e.g. colour or texture
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a kind of dress ornament sample set creation method and device, methods described includes：Obtain the texture information in dress ornament region in sample image；Colour atla collection is set up according to preset color；The dress ornament region is colored according to the colour atla collection and the texture information；According to the dress ornament Area generation dress ornament sample set after coloring in.The program is colored in generate dress ornament sample set by the dress ornament region in the colour atla set pair sample image set up according to preset color, limited sample image obtains substantial amounts of sample set by coloring in for different colour atlas, compared with prior art, overcome Large Copacity sample set in the prior art and be difficult to the defect for obtaining, greatly reduce the workload for obtaining dress ornament sample set, the flexible foundation of colour atla collection, not only enrich the diversity of dress ornament sample set, and realize the automatic expansion of dress ornament sample set, the harmony between the dress ornament sample size of all kinds of colors is ensure that simultaneously.
Description
Technical field
The present invention relates to art of image analysis, and in particular to a kind of dress ornament sample set creation method and device.
Background technology
Pedestrian's dress ornament colour recognition based on deep learning is that intelligent image analysis carries out object feature to monitoring scene
The important component of analysis.In order to be issued to discrimination very high in conditions such as different scenes, light, angles, it is necessary to right
Great amount of samples is trained, while the harmony between sample is kept as far as possible, so video acquisition and sample mark work
It is the link in most cost source in whole development.
Conventional colour recognition has colour atla to contrast with limited self-defined classification (such as：Red, green, blue, it is black, white etc.) side
Method, both of which has the defect of itself.Colour atla method is not particularly suited for pedestrian's dress ornament, because due to light reason, pedestrian's clothes
A homochromy region is adornd and shows as, the colour atla based on statistics defines method and is easy for mistake；And it is based on limited self-defined classification
Method sample collect and mark relatively difficult, while also bad using scalability, such as, demand 1 will recognize pedestrian's dress ornament
Customized N kinds color, demand 2 needs to recognize the color that the other M kinds of pedestrian's dress ornament are made by oneself, then two workloads of demand
It is cumulative, dress ornament sample set needs to re-start mark, and the model training based on deep learning needs substantial amounts of training
Sample, more reduces the operability of the method.Further, since the shades of colour distribution of pedestrian's dress ornament is serious in nature
Unbalanced, such as the dress ornament color such as black, white, grey is more, and the dress ornament color such as purple, green, yellow is just relatively fewer, in a balanced way dress ornament sample
This and quarry be also a problem.
Therefore, the difficulty for obtaining dress ornament sample set, the diversity for increasing dress ornament sample set, optimization dress ornament sample how to be reduced
The harmony of collection, as a technical problem urgently to be resolved hurrily.
The content of the invention
Therefore, the technical problem to be solved in the present invention be obtain in the prior art dress ornament sample set difficulty is big and clothes
The harmony of decorations sample set is poor.
So as to provide a kind of dress ornament sample set creation method and device.
In view of this, the first aspect of the embodiment of the present invention provides a kind of dress ornament sample set creation method, including：Obtain
The texture information in dress ornament region in sample image；Colour atla collection is set up according to preset color；According to the colour atla collection and described
Texture information colors in the dress ornament region；According to the dress ornament Area generation dress ornament sample set after coloring in.
Preferably, it is described colour atla collection is set up according to preset color to include：Color classification is carried out to the preset color；To every
Class color carries out sampling on chromatogram and obtains described per the corresponding multiple colour atlas of class color；By the collection cooperation of the multiple colour atla
It is the colour atla collection.
Preferably, it is described that the bag that colors in is carried out to the dress ornament region according to the colour atla collection and the texture information
Include：The model that colors in based on Poisson's equation is set up according to the colour atla collection and the texture information；To the model that colors in
Solve obtaining the value that colors in；The dress ornament region is colored according to the value that colors in.
Preferably, the texture information for obtaining dress ornament region in sample image includes：Extract the sample image
Dress ornament region；The dress ornament region is carried out into gray processing；The gradient information in the dress ornament region after gray processing is extracted as described
Texture information.
Preferably, also include：Color of object is identified using the dress ornament sample set；Institute cannot correctly identified
When stating color of object, the colour atla of the color of object is added into the colour atla collection, to update the colour atla collection；According to the target
The colour atla of color and the texture information color in the dress ornament region, to update the dress ornament sample set.
The second aspect of the embodiment of the present invention provides a kind of dress ornament sample set generating means, including：Acquisition module, is used for
Obtain the texture information in dress ornament region in sample image；Module is set up, for setting up colour atla collection according to preset color；Color in
Module, for being colored in the dress ornament region according to the colour atla collection and the texture information；Generation module, is used for
According to the dress ornament Area generation dress ornament sample set after coloring in.
Preferably, the module of setting up includes：Taxon, for carrying out color classification to the preset color；Sampling
Module, obtains described per the corresponding multiple colour atlas of class color for carrying out sampling on chromatogram to every class color；Collection modules,
For using the set of the multiple colour atla as the colour atla collection.
Preferably, the module that colors in includes：Unit is set up, for according to the colour atla collection and the texture information
Set up the model that colors in based on Poisson's equation；Unit is solved, for solve obtaining the value that colors in the model that colors in；Color in
Unit, colors in for the value that colored according to the dress ornament region.
Preferably, the acquisition module includes：First extraction unit, the dress ornament region for extracting the sample image；
Gray shade unit, for the dress ornament region to be carried out into gray processing；Second extraction unit, for extracting the dress ornament region after gray processing
Gradient information as the texture information.
Preferably, also include：Identification module, for being identified to color of object using the dress ornament sample set；First
Update module, for when the color of object correctly cannot be identified, the colour atla of the color of object being added into the colour atla
Collection, to generate new colour atla collection；Second update module, for the colour atla according to the color of object and the texture information
The dress ornament region is colored in, to update the dress ornament sample set.
Technical scheme has advantages below：
Dress ornament sample set creation method provided in an embodiment of the present invention and device, by the colour atla set up according to preset color
Dress ornament region in set pair sample image is colored in generate dress ornament sample set, and wherein dress ornament region remains the texture of dress ornament
Structural information, limited sample image obtains substantial amounts of sample set by coloring in for different colour atlas, compared with prior art, gram
Take Large Copacity sample set in the prior art and be difficult to the defect for obtaining, greatly reduce the workload for obtaining dress ornament sample set, color
The flexible foundation of Ka Ji, not only enriches the diversity of dress ornament sample set, and realizes the automatic expansion of dress ornament sample set, together
When ensure that harmony between the dress ornament sample size of all kinds of colors.
Brief description of the drawings
In order to illustrate more clearly of the specific embodiment of the invention or technical scheme of the prior art, below will be to specific
The accompanying drawing to be used needed for implementation method or description of the prior art is briefly described, it should be apparent that, in describing below
Accompanying drawing is some embodiments of the present invention, for those of ordinary skill in the art, before creative work is not paid
Put, other accompanying drawings can also be obtained according to these accompanying drawings.
Fig. 1 is a flow chart of the dress ornament sample set creation method of the embodiment of the present invention 1；
Fig. 2 is one of the pixel at dress ornament region and non-dress ornament zone boundary in the sample image of the embodiment of the present invention 1
Individual schematic diagram；
Fig. 3 is a block diagram of the dress ornament sample set generating means of the embodiment of the present invention 2.
Specific embodiment
Technical scheme is clearly and completely described below in conjunction with accompanying drawing, it is clear that described implementation
Example is a part of embodiment of the invention, rather than whole embodiments.Based on the embodiment in the present invention, ordinary skill
The every other embodiment that personnel are obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
In the description of the invention, it is necessary to illustrate, term " first ", " second " are only used for describing purpose, and can not
It is interpreted as indicating or implying relative importance.
As long as additionally, technical characteristic involved in invention described below different embodiments non-structure each other
Can just be combined with each other into conflict.
Embodiment 1
The present embodiment provides a kind of dress ornament sample set creation method, the sample that can be used for during pedestrian's dress ornament colour recognition
Collection is obtained, as shown in figure 1, comprising the following steps：
S11：Obtain the texture information in dress ornament region in sample image.Sample image can be from different images herein
Or chosen in video material, for different scenes, different gestures, different directions, different sexes, all ages and classes, difference in material
The pedestrian of figure chooses sample image, in order to can as far as possible include the pedestrian's dress ornament information under all situations, so
The texture information in dress ornament region in sample image is extracted afterwards, and texture information is including but not limited to clothes pleat herein
Wrinkle, light and shade region etc..Used as a kind of preferred scheme, step S11 can include：The dress ornament region of sample image, tool are extracted first
Body ground, if sample image is represented with I, can make the mask images of I using related drawing instrument, such as use
The mask images that Photoshop makes I are designated as M, and M marked dress ornament region and background area in sample image, such as mask
Dress ornament region is noted as white in image M, is represented with symbol Ω；Background area is noted as black, is represented with symbol Ο, the back of the body
The region abutted with dress ornament region Ω on scene area Ο is designated asIt is region to set p (i, j)On pixel, then p (i,
J) meetNΩRepresent pixel neighborhood of a point on Ω；Then by dress ornament region
Carry out gray processing, specifically sample image I can be carried out gray processing, will I colouring information removal, retain I texture and
Structural information, it is that a color information for eliminating I remains the texture of I and the gray scale of structural information to obtain gray level image G, G
Image；The gradient information in the dress ornament region after gray processing is finally extracted as texture information, in order to retain the row after tinting
People's dress ornament region keeps the architectural feature of original dress ornament, i.e. clothes fold, light and shade region etc., and dress ornament region must be protected after being colored in
Hold the gradient information of original image.The texture structure in dress ornament region can be characterized using the gradient information in dress ornament region.Such as may be used
To set up size as w × h × 8 matrix ν, for the gradient vector field (gradient information) of storage image G, element in formula w × h × 8
Wide, high, eight Grad in neighborhood direction of each pixel of image G are represented respectively, and the picture on gray level image G is represented with G (i, j)
Vegetarian refreshments, can specifically use equation below calculating matrix：
Then according to above-mentioned formula (1) and dress ornament region, corresponding pixel is calculated dress ornament in gray level image G
The gradient information in region, as the texture information in dress ornament region.
S12：Colour atla collection is set up according to preset color.Depending on preset color is according to concrete application scene herein, for difference
Application scenario different preset colors can be set, and then set up different colour atla collection, it is to avoid single colour atla collection can not be accurate
The problem of dress ornament color is recognized, the flexibility of colour atla collection is improve.Used as a kind of preferred scheme, step S12 can include：To pre-
If color carries out color classification；Every class color is carried out sampling on chromatogram and obtains the corresponding multiple colour atlas of every class color；Will
The set of multiple colour atlas is used as colour atla collection.Specifically, if it is h to need identification preset color classification number, classification number uses l respectivelyh(h
It is nonnegative integer) represent, these colour types are made with correspondence colour atla subset, it is designated asEach colour atla subsetInclude
During multiple colour atlas, such as h=1Red classification is represented, then the red classification can include multiple colour atlas related to red,
On chromatogram, such as (R, G, B) chromatogram, a class color has certain number range, is taken in the number range different
Value can obtain different relevant colors, be sampled by chromatogram (sample frequency can be made by oneself), you can obtainIt is right
The multiple colour atlas answered, i.e.,N isThe number of middle colour atla element, N can use non-negative
Integer, each colour atla element that colour atla is concentrated can be expressed as with three-dimensional vector
S13：Dress ornament region is colored according to colour atla collection and texture information；Herein for each of colour atla concentration
Individual colour atla will color in dress ornament region, i.e., the texture information in dress ornament region passes through in the sample image for once obtaining
After different colour atlas colors in, huge dress ornament sample set can be obtained, the resource of dress ornament sample set is obtained so as to greatly save.Make
It is a kind of preferred scheme, step S13 can include：
Step one：The model that colors in based on Poisson's equation is set up according to colour atla collection and texture information, is mainly adopted here
Colored in principle with discrete Poisson's equation, Poisson's equation can hold image texture structure it is constant in the case of change the face of image
Color, therefore the model that colors in can be set up using equation below (2)：
In above formula (2), the function representation of the pixel value of S certain passage is represented with f；P and q are represented a little and p and q are adjacent each other
Domain, fpRepresent pixel value, f at point pqRepresent the pixel value at point q, then νpqRepresent on two line directions of p q
The value of gradient information ν；Color is representedIn a value of current channel, such as currently processed is that R leads to
Road, then color=r.
Step 2：Then the model that colors in solve and obtain the value that colors in, concrete operation step is as follows：
1：Newly-built sample image S, S are triple channel image, and size is identical with I, and the pixel value of S is all initialized as entirely
Black, i.e. all pixels value is all 0；
2：Select a kind of colour atlaS is colored in, i.e., the color for being shown in the Ω of dress ornament region using S
ForGradient information is equal to ν.Tri- passages of R, G, B seek pixel value respectively, are then combined and obtain colored sample
This image S, three passage processing modes are the same.Solution procedure can be calculated using equation below (3)：
N in above formula (3)pRepresent eight neighborhood directions of p points pixel, hence it is evident that, above formula is to meet the solution to model that colors in, and above formula will
Image dyeing course is described as：It is in boundary line that image S must is fulfilled for initial conditions in the representative function f of dress ornament region Ω
Upper, initial value is color, is the optimization problem of ν in region Ω inside gradients information.Each pixel p above formulas (3) can be represented
Into at least one unknown number, at most 7 equations of unknown number, as shown in Fig. 2 the p that sets up an officeijRepresent the point in dress ornament region, qijTable
Show dress ornament edges of regions point, i.e. pij∈ Ω,Enumerate by formula (3) with p22And p23Centered on put equation (4) and
Formula (5) is as follows：
Formula (4) and formula (5) left side are all unknown numbers, and the right is all datum, are all typical once linear equations, each
Formula (3) can be launched into a similar formula (4) and once linear equation as (5) by the point in individual dress ornament region Ω, if setting
Size (Ω)=k, then just have k such set of equations into a system of linear equations, equation group unknown number number is also
k.Formula (4) and formula (5) can be solved using Gauss-Seidel iterative methods, obtain the value that colors in.
Step 3：Finally according to coloring in, value colors in dress ornament region, and the value that will color in fills out image by the position of point p
In S, that is, complete the task that colors in the single passages of image S.By the way of step 2, respectively to three passages of image S
Solve and obtain the corresponding value that colors in, and carry out coloring in for related channel program, be then combined with together, that is, complete in dress ornament area
Image S is colored in the Ω of domain, then allow image S non-Ω regions with artwork I as, that is, keep the background of artwork, it is finally complete
Dress ornament region carries out the work of designated color dyeing in a sample image into I.Similarly, traversal colour atla is concentrated
Colour atla colors in the dress ornament region of all sample images in I.
S14：According to the dress ornament Area generation dress ornament sample set after coloring in.The colour atla that traversal colour atla is concentrated is to all samples in I
After the dress ornament region of this image is colored in, you can obtain huge dress ornament sample set, so manually need not on the spot select mesh
Mark the dress ornament sample of color, you can automatically generate the dress ornament sample set of substantial amounts of different colours, not only operation difficulty reduction, and
Workload is greatly reduced, while ensure that the harmony between all kinds of color cards.
As a kind of preferred scheme, the step of also expansion including sample：Color of object is known using dress ornament sample set
Not；When color of object correctly cannot be identified, the colour atla of color of object is added into colour atla collection, to update colour atla collection；According to mesh
The colour atla and texture information for marking color color in dress ornament region, to update dress ornament sample set.Specifically, step can be used
The dress ornament sample set obtained in rapid S14 is trained to dress ornament identification model, then chooses color of object and recognizes mould to the dress ornament
Type is tested, if identification mistake, illustrates, the color of object may not be in dress ornament sample set, then can be by the color of object
Corresponding colour atla is added in the corresponding colour atla subset of relevant colors classification, completes the expansion of colour atla collection, and use the target
The colour atla of color to the dress ornament region of each image in sample image I color in and obtains a new dress ornament sample, and
The new dress ornament sample is added to dress ornament sample set, the automatic expansion to dress ornament sample set is realized, can be used and step herein
Rapid S13 the same day colors in scheme, will not be repeated here.
The dress ornament sample set creation method that the present embodiment is provided, by the colour atla set pair sample graph set up according to preset color
Dress ornament region as in is colored in generate dress ornament sample set, and wherein dress ornament region remains the texture information of dress ornament,
Limited sample image obtains substantial amounts of sample set by coloring in for different colour atlas, compared with prior art, overcomes existing
Large Copacity sample set is difficult to the defect for obtaining in technology, greatly reduces the workload for obtaining dress ornament sample set, the spirit of colour atla collection
It is living to set up, the diversity of dress ornament sample set is not only enriched, and the automatic expansion of dress ornament sample set is realized, while ensure that
Harmony between the dress ornament sample size of all kinds of colors.
Embodiment 2
The present embodiment has supplied a kind of dress ornament sample set generating means, the sample that can be used for during pedestrian's dress ornament colour recognition
Collection is obtained, as shown in figure 3, including：Acquisition module 31, set up module 32, the module that colors in 33 and generation module 34, each functions of modules
It is as follows：
Acquisition module 31, the texture information for obtaining dress ornament region in sample image, referring specifically in embodiment 1
To the detailed description of step S11.
Set up module 32, for setting up colour atla collection according to preset color, referring specifically in embodiment 1 to the detailed of step S12
Thin description.
The module that colors in 33, for being colored in dress ornament region according to colour atla collection and texture information, referring specifically to reality
Apply the detailed description to step S13 in example 1.
Generation module 34, for according to the dress ornament Area generation dress ornament sample set after coloring in.Referring specifically to right in embodiment 1
The detailed description of step S14.
Used as a kind of preferred scheme, setting up module 32 includes：Taxon, for carrying out color classification to preset color；
Sampling module, the corresponding multiple colour atlas of every class color are obtained for carrying out sampling on chromatogram to every class color；Collection modules,
For using the set of multiple colour atlas as colour atla collection.Retouched referring specifically to the detailed of preferred scheme in embodiment 1 to step S12
State.
Used as a kind of preferred scheme, the module that colors in 33 includes：Unit is set up, for according to colour atla collection and texture information
Set up the model that colors in based on Poisson's equation；Unit is solved, for solve obtaining the value that colors in the model that colors in；Color in list
Unit, for being colored in dress ornament region according to the value that colors in.Referring specifically in embodiment 1 to the detailed of the preferred scheme of step S13
Thin description.
Used as a kind of preferred scheme, acquisition module 31 includes：First extraction unit, the dress ornament area for extracting sample image
Domain；Gray shade unit, for dress ornament region to be carried out into gray processing；Second extraction unit, for extracting the dress ornament region after gray processing
Gradient information as texture information.Referring specifically in embodiment 1 to the detailed description of the preferred scheme of step S11.
As a kind of preferred scheme, also include：Identification module, for being known to color of object using dress ornament sample set
Not；First update module, for when color of object correctly cannot be identified, the colour atla of color of object being added into colour atla collection, with
The new colour atla collection of generation；Second update module, enters for the colour atla and texture information according to color of object to dress ornament region
Row colors in, to update dress ornament sample set.Described in detail referring specifically to the correlation in embodiment 1.
The dress ornament sample set generating means that the present embodiment is provided, by the colour atla set pair sample graph set up according to preset color
Dress ornament region as in is colored in generate dress ornament sample set, and wherein dress ornament region remains the texture information of dress ornament,
Limited sample image obtains substantial amounts of sample set by coloring in for different colour atlas, compared with prior art, overcomes existing
Large Copacity sample set is difficult to the defect for obtaining in technology, greatly reduces the workload for obtaining dress ornament sample set, the spirit of colour atla collection
It is living to set up, the diversity of dress ornament sample set is not only enriched, and the automatic expansion of dress ornament sample set is realized, while ensure that
Harmony between the dress ornament sample size of all kinds of colors.
Obviously, above-described embodiment is only intended to clearly illustrate example, and not to the restriction of implementation method.It is right
For those of ordinary skill in the art, can also make on the basis of the above description other multi-forms change or
Change.There is no need and unable to be exhaustive to all of implementation method.And the obvious change thus extended out or
Among changing still in the protection domain of the invention.
Claims (10)
1. a kind of dress ornament sample set creation method, it is characterised in that including：
Obtain the texture information in dress ornament region in sample image；
Colour atla collection is set up according to preset color；
The dress ornament region is colored according to the colour atla collection and the texture information；
According to the dress ornament Area generation dress ornament sample set after coloring in.
2. dress ornament sample set creation method according to claim 1, it is characterised in that described that color is set up according to preset color
Ka Ji includes：
Color classification is carried out to the preset color；
Every class color is carried out sampling on chromatogram, and it is described per the corresponding multiple colour atlas of class color to obtain；
Using the set of the multiple colour atla as the colour atla collection.
3. dress ornament sample set creation method according to claim 1, it is characterised in that described according to the colour atla collection and institute
State texture information the dress ornament region is colored in including：
The model that colors in based on Poisson's equation is set up according to the colour atla collection and the texture information；
The model that colors in solve obtaining the value that colors in；
The dress ornament region is colored according to the value that colors in.
4. dress ornament sample set creation method according to claim 1, it is characterised in that dress ornament in the acquisition sample image
The texture information in region includes：
Extract the dress ornament region of the sample image；
The dress ornament region is carried out into gray processing；
The gradient information in the dress ornament region after gray processing is extracted as the texture information.
5. dress ornament sample set creation method according to claim 1, it is characterised in that also include：
Color of object is identified using the dress ornament sample set；
When the color of object correctly cannot be identified, the colour atla of the color of object is added into the colour atla collection, to update
The colour atla collection；
Colour atla and the texture information according to the color of object color in the dress ornament region, described to update
Dress ornament sample set.
6. a kind of dress ornament sample set generating means, it is characterised in that including：
Acquisition module, the texture information for obtaining dress ornament region in sample image；
Module is set up, for setting up colour atla collection according to preset color；
Color in module, for being colored in the dress ornament region according to the colour atla collection and the texture information；
Generation module, for according to the dress ornament Area generation dress ornament sample set after coloring in.
7. dress ornament sample set generating means according to claim 6, it is characterised in that the module of setting up includes：
Taxon, for carrying out color classification to the preset color；
Sampling module, obtains described per the corresponding multiple colour atlas of class color for carrying out sampling on chromatogram to every class color；
Collection modules, for using the set of the multiple colour atla as the colour atla collection.
8. dress ornament sample set generating means according to claim 6, it is characterised in that the module that colors in includes：
Unit is set up, for setting up the model that colors in based on Poisson's equation according to the colour atla collection and the texture information；
Unit is solved, for solve obtaining the value that colors in the model that colors in；
Color in unit, and the dress ornament region is colored in for the value that colored according to.
9. dress ornament sample set generating means according to claim 6, it is characterised in that the acquisition module includes：
First extraction unit, the dress ornament region for extracting the sample image；
Gray shade unit, for the dress ornament region to be carried out into gray processing；
Second extraction unit, for extracting the gradient information in the dress ornament region after gray processing as the texture information.
10. dress ornament sample set generating means according to claim 6, it is characterised in that also include：
Identification module, for being identified to color of object using the dress ornament sample set；
First update module, for when the color of object correctly cannot be identified, the colour atla of the color of object being added
The colour atla collection, to generate new colour atla collection；
Second update module, enters for the colour atla according to the color of object and the texture information to the dress ornament region
Row colors in, to update the dress ornament sample set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710038546.XA CN106898035A (en) | 2017-01-19 | 2017-01-19 | A kind of dress ornament sample set creation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710038546.XA CN106898035A (en) | 2017-01-19 | 2017-01-19 | A kind of dress ornament sample set creation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106898035A true CN106898035A (en) | 2017-06-27 |
Family
ID=59198245
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710038546.XA Pending CN106898035A (en) | 2017-01-19 | 2017-01-19 | A kind of dress ornament sample set creation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106898035A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111062862A (en) * | 2019-12-19 | 2020-04-24 | 北京澎思科技有限公司 | Color-based data enhancement method and system, computer device and storage medium |
| CN113537410A (en) * | 2021-09-14 | 2021-10-22 | 征图新视(江苏)科技股份有限公司 | Universal automatic balancing method for deep learning positive samples |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101677354A (en) * | 2008-09-16 | 2010-03-24 | 富士施乐株式会社 | Image forming system, image forming apparatus, terminal device, gradation characteristic measuring method |
| CN102915452A (en) * | 2012-10-12 | 2013-02-06 | 广东欧珀移动通信有限公司 | Method and device for acquiring color card number corresponding to color |
| CN104063888A (en) * | 2014-06-19 | 2014-09-24 | 云南大学 | Pop art style drawing method based on non-photorealistic |
| CN104899389A (en) * | 2015-06-17 | 2015-09-09 | 张梅云 | Intelligent costume design method |
| CN105117935A (en) * | 2015-08-03 | 2015-12-02 | 武汉中研福世科技有限公司 | Realization method of intelligent dress matching and apparatus thereof |
| CN105810209A (en) * | 2016-01-04 | 2016-07-27 | 邱子皓 | Data conversion method based on mapping relation |
-
2017
- 2017-01-19 CN CN201710038546.XA patent/CN106898035A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101677354A (en) * | 2008-09-16 | 2010-03-24 | 富士施乐株式会社 | Image forming system, image forming apparatus, terminal device, gradation characteristic measuring method |
| CN102915452A (en) * | 2012-10-12 | 2013-02-06 | 广东欧珀移动通信有限公司 | Method and device for acquiring color card number corresponding to color |
| CN104063888A (en) * | 2014-06-19 | 2014-09-24 | 云南大学 | Pop art style drawing method based on non-photorealistic |
| CN104899389A (en) * | 2015-06-17 | 2015-09-09 | 张梅云 | Intelligent costume design method |
| CN105117935A (en) * | 2015-08-03 | 2015-12-02 | 武汉中研福世科技有限公司 | Realization method of intelligent dress matching and apparatus thereof |
| CN105810209A (en) * | 2016-01-04 | 2016-07-27 | 邱子皓 | Data conversion method based on mapping relation |
Non-Patent Citations (1)
| Title |
|---|
| 王敏: "海量图像的描述与合成方法研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111062862A (en) * | 2019-12-19 | 2020-04-24 | 北京澎思科技有限公司 | Color-based data enhancement method and system, computer device and storage medium |
| CN113537410A (en) * | 2021-09-14 | 2021-10-22 | 征图新视(江苏)科技股份有限公司 | Universal automatic balancing method for deep learning positive samples |
| CN113537410B (en) * | 2021-09-14 | 2021-12-07 | 征图新视(江苏)科技股份有限公司 | Universal automatic balancing method for deep learning positive samples |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830912B (en) | An Interactive Grayscale Image Colorization Method for Adversarial Learning of Deep Features | |
| CN106296666B (en) | A kind of color image removes shadow method and application | |
| CN105069746B (en) | Video real-time face replacement method and its system based on local affine invariant and color transfer technology | |
| CN109033998A (en) | Remote sensing image atural object mask method based on attention mechanism convolutional neural networks | |
| CN106485288B (en) | An automatic identification method for colored fabric tissue | |
| CN106548169B (en) | Fuzzy literal Enhancement Method and device based on deep neural network | |
| CN101213576A (en) | Album creating apparatus, album creating method and program therefor | |
| CN109920012A (en) | Image coloring system and method based on convolutional neural network | |
| CN108846869B (en) | An automatic color matching method for clothing based on natural image color | |
| CN109657612A (en) | A kind of quality-ordered system and its application method based on facial image feature | |
| CN109191428A (en) | Full-reference image quality evaluating method based on masking textural characteristics | |
| CN108171715A (en) | A kind of image partition method and device | |
| CN104636759A (en) | Method for obtaining picture recommending filter information and picture filter information recommending system | |
| CN106097833A (en) | A kind of Educational toy external member and digit recognition method thereof | |
| CN109583349A (en) | A kind of method and system for being identified in color of the true environment to target vehicle | |
| CN110298893A (en) | A kind of pedestrian wears the generation method and device of color identification model clothes | |
| EP1302896A1 (en) | Process of identification of sky in an image and image obtained using this process | |
| CN107045722B (en) | Merge the video signal process method of static information and multidate information | |
| CN106898035A (en) | A kind of dress ornament sample set creation method and device | |
| CN112287838A (en) | Cloud and fog automatic identification method and system based on static meteorological satellite image sequence | |
| CN105223208A (en) | A kind of circuit board detecting template and preparation method thereof, circuit board detecting method | |
| Du et al. | Double-channel guided generative adversarial network for image colorization | |
| CN106558044A (en) | The resolution measuring method of image module | |
| CN105812668A (en) | Image processing method, processing device and photographing apparatus | |
| CN106652046A (en) | Model tinting method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170627 |
|
| RJ01 | Rejection of invention patent application after publication |