CN107909580A - A kind of pedestrian wears color identification method, electronic equipment and storage medium clothes - Google Patents
A kind of pedestrian wears color identification method, electronic equipment and storage medium clothes Download PDFInfo
- Publication number
- CN107909580A CN107909580A CN201711059617.0A CN201711059617A CN107909580A CN 107909580 A CN107909580 A CN 107909580A CN 201711059617 A CN201711059617 A CN 201711059617A CN 107909580 A CN107909580 A CN 107909580A
- Authority
- CN
- China
- Prior art keywords
- picture
- segmentation
- pedestrian
- color
- feature
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of pedestrian to wear color identification method clothes, including：Obtain original pedestrian's picture；Region segmentation is carried out to original pedestrian's picture using background segment model；Segmentation picture is obtained according to the result of region segmentation, segmentation picture includes segmentation picture above the waist and lower part of the body segmentation picture；Pedestrian in identification segmentation picture wears color clothes.The invention further relates to a kind of electronic equipment and storage medium, pedestrian provided by the invention wears color identification method, electronic equipment and storage medium clothes and carries out region segmentation to original pedestrian's picture, segmentation picture above the waist and lower part of the body segmentation picture are obtained according to the result of region segmentation, reduce background interference；Colour recognition is carried out respectively to upper part of the body segmentation picture and lower part of the body segmentation picture, improves colour recognition effect.
Description
Technical field
The present invention relates to image and colour recognition field, more particularly to a kind of pedestrian wears color identification method clothes, electronics is set
Standby and storage medium.
Background technology
Colour recognition is a traditional problem, and many important applications, many tradition have also been expedited the emergence of out based on colour recognition
Color characteristic such as RGB or hsv color histogram feature be widely used in all trades and professions；Such as the identification of vehicle color,
In traditional Car license recognition License Plate is carried out as subsidiary conditions by color characteristic；In some industrial circles by color
Identify to distinguish object, such as the sorting system of colored pencil production line, point of different colours pencil is carried out with color characteristic
Pick, the simple object identification of some backgrounds also there are many methods for having used colour recognition；Colour recognition is in safety-security area
Have important application, accurate colour recognition, greatly promotes and looks for people's efficiency, pedestrian wear clothes colour recognition be intended to identification video or
The clothes of pedestrian and trousers color in image.
Having many pedestrians to wear colour recognition clothes at present is stated using conventional color feature, special by the color for extracting sample
Sign, the mode of one grader of retraining realize, for example, using HSV space statistical color histogram method respectively to pedestrian
Jacket and trousers be identified：First in tri- passages of RGB with Sobel operators calculated level and vertical gradient projection from
And the cut-off rule of the lower part of the body is obtained, then determine cut-off rule by artificially limiting the position section of cut-off rule, hsv color is empty
Between be divided into 8 color regions, by color of image space from RGB be transformed into HSV after statistics fall 8 color regions pixel
Number, so as to obtain clothing color.For example use PETA data sets：Various pedestrian's attributes are contained in data set, including
Above the waist and lower part of the body color attribute, the method identifying rows using SVM and markov random file are humanized, or using opening
Set of source data PETA carries out network training and test.Both of which is that the upper lower part of the body color labeled data that make use of mark is instructed
Practice grader, Attribute Recognition is carried out using the grader after training.
Traditional pedestrian wears color identification method clothes and carries out colour recognition, method using conventional color histogram statistical features
Simply, it is widely used, recognition effect is relatively good when application scenarios are simple, background is single, illumination effect is small, but on the scene
Recognition effect is very poor in the case that scape is complicated, illumination effect is larger.In monitoring scene, complex background, block, light the problems such as
Influence it is very serious, it is larger that the methods of directly using conventional color histogram under such circumstances, extracts feature interference, obtains
Recognition effect is very poor after feature.
The content of the invention
For overcome the deficiencies in the prior art, it is an object of the present invention to provide a kind of pedestrian to wear colour recognition side clothes
Method, to solve the problems, such as that it is poor that existing pedestrian wears recognition effect in colour recognition clothes.
The second object of the present invention is to provide a kind of electronic equipment, to solve to identify in existing pedestrian's clothing colour recognition
The problem of effect is poor.
An object of the present invention adopts the following technical scheme that realization：
A kind of pedestrian wears color identification method clothes, including：
Obtain original pedestrian's picture；
Region segmentation is carried out to original pedestrian's picture using background segment model, wherein, the background segment model
Training be based on deep neural network；
Obtained splitting picture according to the result of the region segmentation, the segmentation picture includes splitting above the waist picture with
Half body splits picture；
Identify that the pedestrian in the segmentation picture wears color clothes.
Further, it is described that original pedestrian's picture progress region segmentation is included using background segment model：
Convolutional calculation and normalized are carried out to the original image to extract original image feature；
To be background segment feature by normalized original image Feature Conversion；
Region segmentation information is obtained according to the background segment feature.
Further, the pedestrian in the identification segmentation picture, which wears color clothes, includes：
Convolutional calculation is carried out to the segmentation picture to extract segmentation picture feature；
Nonlinear transformation is carried out to the segmentation picture feature；
Dimension-reduction treatment is carried out to the segmentation picture feature Jing Guo nonlinear transformation；
Linear transformation is carried out to the segmentation picture feature Jing Guo dimension-reduction treatment to obtain characteristic color；
Identify that the pedestrian in segmentation picture wears color clothes according to the characteristic color.
Further, the pedestrian identified according to the characteristic color in segmentation picture, which wears color clothes, includes：
Judge that the pedestrian in the segmentation picture wears whether color is pure color clothes；
If it is that the colour type of the pure color is identified for pure color that the pedestrian in the segmentation picture, which wears color clothes,.
Further, the result according to the region segmentation, which obtains segmentation picture, includes：
Upper part of the body profile and lower part of the body profile are obtained according to the result of the region segmentation；
Segmentation picture and the lower part of the body above the waist are extracted according to the upper part of the body profile and the lower part of the body profile respectively
Split picture.
Further, it is described that segmentation picture above the waist is extracted according to the upper part of the body profile and the lower part of the body profile respectively
Include with lower part of the body segmentation picture：
Upper part of the body profile coordinate and lower part of the body profile coordinate are calculated according to the upper part of the body profile and the lower part of the body profile；
The outer of the upper part of the body profile is extracted according to the upper part of the body profile coordinate and the lower part of the body profile coordinate respectively
Connect the boundary rectangle of rectangle and the lower part of the body profile.
The second object of the present invention adopts the following technical scheme that realization：
A kind of electronic equipment, including：Processor；
Memory；And program, wherein described program is stored in the memory, and is configured to by processor
Perform, described program includes being used to perform：
Obtain original pedestrian's picture；
Region segmentation is carried out to original pedestrian's picture using background segment model, wherein, the background segment model
Training be based on deep neural network；
Obtained splitting picture according to the result of the region segmentation, the segmentation picture includes splitting above the waist picture with
Half body splits picture；
Identify that the pedestrian in the segmentation picture wears color clothes.
Further, described program is additionally operable to perform：
Convolutional calculation and normalized are carried out to the original image to extract original image feature；
To be background segment feature by normalized original image Feature Conversion；
Region segmentation information is obtained according to the background segment feature.
Further, described program is additionally operable to perform：
Convolutional calculation is carried out to the segmentation picture to extract segmentation picture feature；
Nonlinear transformation is carried out to the segmentation picture feature；
Dimension-reduction treatment is carried out to the segmentation picture feature Jing Guo nonlinear transformation；
Linear transformation is carried out to the segmentation picture feature Jing Guo dimension-reduction treatment to obtain characteristic color；
Identify that the pedestrian in segmentation picture wears color clothes according to the characteristic color.
The invention further relates to a kind of computer-readable recording medium, is stored thereon with computer program, the computer journey
Sequence is executed by processor above-mentioned method.
Compared with prior art, the beneficial effects of the present invention are：Region segmentation is carried out to original pedestrian's picture, according to region
The result of segmentation obtains segmentation picture above the waist and lower part of the body segmentation picture, reduces background interference；To the upper part of the body segmentation picture and
Lower part of the body segmentation picture carries out colour recognition respectively, improves colour recognition effect.
Brief description of the drawings
Fig. 1 wears color identification method flow chart clothes for pedestrian provided in an embodiment of the present invention；
Fig. 2 is the schematic diagram of electronic equipment provided in an embodiment of the present invention.
Embodiment
In the following, with reference to attached drawing and embodiment, the present invention is described further, it is necessary to which explanation is, not
Under the premise of afoul, new implementation can be formed between various embodiments described below or between each technical characteristic in any combination
Example.
As shown in Figure 1, pedestrian provided in an embodiment of the present invention wears color identification method clothes, including：
Step S101：Obtain original pedestrian's picture.
Specifically, original pedestrian's picture to be identified is obtained from video.
Step S102：Region segmentation is carried out to original pedestrian's picture using background segment model, wherein, the background
The training of parted pattern is based on deep neural network.
The training for the background segment model that the embodiment of the present invention uses is based on deep neural network, training background segment model
Sample data be pedestrian's picture, treat trained pedestrian's picture and mark the pixels such as the upper part of the body, the lower part of the body, knapsack, shoes respectively
Region, different pixel regions are identified using different colors, by by the upper part of the body of pedestrian, the lower part of the body, and equipment, footwear
Son, whole body etc. are learnt as single passage.Such as：When to learning above the waist, the area label of the upper part of the body
For 1, other regions are 0, and deep neural network only splits upper part of the body region to upper part of the body passage, similarly, to other passages only
Study segmentation corresponding region.The deep neural network structure include convolutional layer, pond layer, batch normalize layer, Feature Conversion layer with
And warp lamination, wherein convolutional layer make picture convolution operation, extract picture feature；The picture feature that pond layer obtains convolution
Dimensionality reduction is carried out, picture feature is normalized in batch normalization layer, so as to simplify training, accelerates model convergence；Feature turns
The feature that common full figure feature is converted into background segment needs by layer by mapping operations is changed, training deep neural network, obtains
To background segment model.
The step includes：
Step S1021：Convolutional calculation and normalized are carried out to the original image to extract original image feature；
Specifically, original image is inputted background segment model, convolutional layer carries out convolutional calculation to extract to original image
Original image feature, pond layer carry out dimension-reduction treatment to original image feature, extract the main feature of original image feature, often
There is batch normalization layer behind one convolutional layer, the main feature of original image is normalized in batch normalization layer, with simplification
Calculating process.
Step S1022：To be background segment feature by normalized original image Feature Conversion.
Specifically, background segment feature will be obtained by mapping operations by normalized main feature.
Step S1023：Region segmentation information is obtained according to the background segment feature.
Specifically, according to the corresponding area label information of background segment feature, confirm region segmentation information, identify original
The region such as pedestrian's upper part of the body, the lower part of the body, equipment, shoes in picture.
Step S103：Segmentation picture is obtained according to the result of the region segmentation, the segmentation picture includes upper half status
Cut picture and lower part of the body segmentation picture.
The step includes：
Step S1031：Upper part of the body profile and lower part of the body profile are obtained according to the result of the region segmentation.
Specifically, original image after background segment model treatment, obtains original from obtained region segmentation information
The upper part of the body profile and lower part of the body profile of color to be identified in picture.The dress of pedestrian can also be identified according to colour recognition requirement
Standby profile or shoes profile, only to be illustrated the problem of identifying upper part of the body color and lower part of the body color in the present embodiment.
Step S1032：According to the upper part of the body profile and the lower part of the body profile extract respectively above the waist segmentation picture and
The lower part of the body splits picture.
Specifically, upper part of the body profile coordinate and lower part of the body profile coordinate are calculated according to upper part of the body profile and lower part of the body profile；
Confirm the outermost coordinate of upper part of the body profile and the outermost coordinate of lower part of the body profile, extract the external square of upper part of the body profile respectively
The boundary rectangle of shape and lower part of the body profile, exports the picture having powerful connections.The picture filled relative to black picture element, using having powerful connections
Picture recognition color effects it is more preferable.
Step S104：Identify that the pedestrian in the segmentation picture wears color clothes.
Wherein, the colour recognition in the segmentation picture of the embodiment of the present invention is the colour recognition mould based on deep neural network
Type, the sample data of training colour recognition model is the picture by region segmentation, and color task includes 12 kinds of colors altogether,
Red, orange, yellow, green, blue, blue, purple respectively, powder, ash, black, white, brown, in order to avoid autgmentability it is poor the problem of, will above the waist and under
The identical color data of half body trains colour recognition model as identical category.The structure of the neural network model is inputted by data
Layer, convolutional layer, batch standardization layer, non-linear layer, pond layer, full articulamentum composition, finally classifies plus softmax, defeated
Enter data input layer can carry out picture pretreatment operation, such as mirror image for the picture Jing Guo region segmentation, the layer, at random
The operation such as cut out, so as to increase training samples number.The quantity of convolutional layer is multiple, will pass through the processed region of input layer point
Cut picture and input first convolutional layer, extract characteristics of image；Convolutional layer below continues to extract picture feature, all convolutional layers
Output all band batch standardization layers, the picture feature that batch standardization layer exports convolutional layer is normalized, so as to optimize
Training, accelerates convergence rate.Non-linear layer carries out the picture feature by convolution and normalized by nonlinear function
Nonlinear transformation so that its feature exported has stronger ability to express.Pond layer is to the picture feature Jing Guo nonlinear transformation
Many-to-one map operation is carried out, further strengthens the non-linear of feature, output characteristic can also be reduced by being handled by pond layer
Size, so as to reduce network parameter.Full articulamentum is to do linear transformation to the picture feature of input, by the Feature Mapping of study
The feature needed to colour recognition.Finally pass through multiple softmax depletion layers, calculate prediction classification and the other error of tag class,
Train colour recognition model.
Specifically, the step includes：
Step S1041：Convolutional calculation is carried out to the segmentation picture to extract segmentation picture feature.
Specifically, will segmentation picture input colour recognition model, the first convolutional calculation by convolutional layer and batch standardization
The normalized of layer, extracts segmentation picture feature.
Step S1042：Nonlinear transformation is carried out to the segmentation picture feature.
Specifically, nonlinear transformation layer carries out nonlinear function computing to segmentation picture feature so that its picture exported
Feature has stronger ability to express.
Step S1043：Dimension-reduction treatment is carried out to the segmentation picture feature Jing Guo nonlinear transformation.
Specifically, pond layer carries out the picture feature Jing Guo nonlinear change many-to-one map operation, it is further strong
Change the non-linear of segmentation picture feature, reduce the size of output characteristic, so as to reduce network parameter.
Step S1044：Linear transformation is carried out to the segmentation picture feature Jing Guo dimension-reduction treatment to obtain characteristic color.
Specifically, full articulamentum carries out linear transformation to the segmentation picture feature handled by pond layer, picture will be split
The feature that Feature Mapping is needed to colour recognition.
Step S1045：Identify that the pedestrian in segmentation picture wears color clothes according to the characteristic color.
Specifically, confirm that the pedestrian in segmentation picture wears color clothes according to the corresponding color of characteristic color.
As preferred embodiment, a branch using polychrome as color, during colour recognition model training,
Pure color, polychrome identification model are first obtained, then increases pure color model branch in pure color, polychrome identification model.Colour recognition process
In, the pedestrian that colour recognition model first judges to split in picture wears whether color is pure color clothes；If identifying face is worn clothes for pedestrian
Color is polychrome, then exports polychrome recognition result, no longer polychrome is finely divided, so as to reduce the probability of misrecognition to polychrome；If
It is that the colour type of pure color is then identified for pure color that pedestrian in segmentation picture, which wears color clothes,.
Pedestrian provided by the invention wears color identification method clothes and is based on deep neural network to original pedestrian's picture progress area
Regional partition, obtains segmentation picture above the waist according to the result of region segmentation and the lower part of the body splits picture, reduce background interference；It is based on
Deep neural network carries out colour recognition respectively to upper part of the body segmentation picture and lower part of the body segmentation picture, improves colour recognition effect
Fruit.
As shown in Fig. 2, electronic equipment provided in an embodiment of the present invention, including：Processor 11, memory 12 and program,
Its Program is stored in memory 12, and is configured to be performed by processor 11, and program includes being used to perform：
Obtain original pedestrian's picture；
Region segmentation is carried out to original pedestrian's picture using background segment model, wherein, the background segment model
Training be based on deep neural network；
Obtained splitting picture according to the result of the region segmentation, the segmentation picture includes splitting above the waist picture with
Half body splits picture；
Identify that the pedestrian in the segmentation picture wears color clothes.
Further, described program is additionally operable to perform：
Convolutional calculation and normalized are carried out to the original image to extract original image feature；
To be background segment feature by normalized original image Feature Conversion；
Region segmentation information is obtained according to the background segment feature.
Further, described program is additionally operable to perform：
To it is described segmentation picture pre-processed and convolutional calculation with obtain segmentation picture feature；
Nonlinear transformation is carried out to the segmentation picture feature；
Dimension-reduction treatment is carried out to the segmentation picture feature Jing Guo nonlinear transformation；
Linear transformation is carried out to the segmentation picture feature Jing Guo dimension-reduction treatment to obtain characteristic color；
Identify that the pedestrian in segmentation picture wears color clothes according to the characteristic color.
The method in electronic equipment and previous embodiment in the present embodiment is based on two sides under same inventive concept
Face, is above being described in detail method implementation process, so those skilled in the art can be according to described above clear
Understand to Chu the implementation process of the electronic equipment in the present embodiment, in order to illustrate the succinct of book, just repeat no more herein.
As seen through the above description of the embodiments, those skilled in the art can be understood that the present invention can
Realized by the mode of software plus required general hardware platform.Based on such understanding, technical scheme essence
On the part that contributes in other words to the prior art can be embodied in the form of software product.The invention further relates to one kind
Computer-readable recording medium, such as ROM/RAM, magnetic disc, CD, are stored thereon with computer program, and computer program is located
Manage device and perform above-mentioned method.
Pedestrian provided by the invention wears color identification method, electronic equipment and storage medium clothes and original pedestrian's picture is carried out
Region segmentation, obtains segmentation picture above the waist according to the result of region segmentation and the lower part of the body splits picture, reduce background interference；It is right
Segmentation picture and lower part of the body segmentation picture carry out colour recognition respectively above the waist, improve colour recognition effect.
The above embodiment is only the preferred embodiment of the present invention, it is impossible to the scope of protection of the invention is limited with this,
The change and replacement for any unsubstantiality that those skilled in the art is done on the basis of the present invention belong to institute of the present invention
Claimed scope.
Claims (10)
1. a kind of pedestrian wears color identification method clothes, it is characterised in that including：
Obtain original pedestrian's picture；
Region segmentation is carried out to original pedestrian's picture using background segment model, wherein, the instruction of the background segment model
White silk is based on deep neural network；
Segmentation picture is obtained according to the result of the region segmentation, the segmentation picture includes segmentation picture and the lower part of the body above the waist
Split picture；
Identify that the pedestrian in the segmentation picture wears color clothes.
2. pedestrian according to claim 1 wears color identification method clothes, it is characterised in that described to use background segment model
Carrying out region segmentation to original pedestrian's picture includes：
Convolutional calculation and normalized are carried out to the original image to extract original image feature；
To be background segment feature by normalized original image Feature Conversion；
Region segmentation information is obtained according to the background segment feature.
3. pedestrian according to claim 2 wears color identification method clothes, it is characterised in that the identification segmentation picture
In pedestrian wear clothes color include：
Convolutional calculation is carried out to the segmentation picture to extract segmentation picture feature；
Nonlinear transformation is carried out to the segmentation picture feature；
Dimension-reduction treatment is carried out to the segmentation picture feature Jing Guo nonlinear transformation；
Linear transformation is carried out to the segmentation picture feature Jing Guo dimension-reduction treatment to obtain characteristic color；
Identify that the pedestrian in segmentation picture wears color clothes according to the characteristic color.
4. pedestrian according to claim 3 wears color identification method clothes, it is characterised in that described according to the colour recognition
Feature recognition, which goes out the pedestrian split in picture, which wears color clothes, includes：
Judge that the pedestrian in the segmentation picture wears whether color is pure color clothes；
If it is that the colour type of the pure color is identified for pure color that the pedestrian in the segmentation picture, which wears color clothes,.
5. pedestrian according to claim 1 wears color identification method clothes, it is characterised in that described according to the region segmentation
Result obtain segmentation picture include：
Upper part of the body profile and lower part of the body profile are obtained according to the result of the region segmentation；
Segmentation picture above the waist and lower part of the body segmentation are extracted according to the upper part of the body profile and the lower part of the body profile respectively
Picture.
6. pedestrian according to claim 5 wears color identification method clothes, it is characterised in that described according to the upper part of the body wheel
Wide and described lower part of the body profile extracts segmentation picture above the waist and lower part of the body segmentation picture respectively to be included：
Upper part of the body profile coordinate and lower part of the body profile coordinate are calculated according to the upper part of the body profile and the lower part of the body profile；
Extract the external square of the upper part of the body profile respectively according to the upper part of the body profile coordinate and the lower part of the body profile coordinate
The boundary rectangle of shape and the lower part of the body profile.
7. a kind of electronic equipment, it is characterised in that including：Processor；
Memory；And program, wherein described program is stored in the memory, and is configured to be held by processor
OK, described program includes being used to perform：
Obtain original pedestrian's picture；
Region segmentation is carried out to original pedestrian's picture using background segment model, wherein, the instruction of the background segment model
White silk is based on deep neural network；
Segmentation picture is obtained according to the result of the region segmentation, the segmentation picture includes segmentation picture and the lower part of the body above the waist
Split picture；
Identify that the pedestrian in the segmentation picture wears color clothes.
8. electronic equipment according to claim 7, it is characterised in that described program is additionally operable to perform：
Convolutional calculation and normalized are carried out to the original image to extract original image feature；
To be background segment feature by normalized original image Feature Conversion；
Region segmentation information is obtained according to the background segment feature.
9. electronic equipment according to claim 8, it is characterised in that described program is additionally operable to perform：
Convolutional calculation is carried out to the segmentation picture to extract segmentation picture feature；
Nonlinear transformation is carried out to the segmentation picture feature；
Dimension-reduction treatment is carried out to the segmentation picture feature Jing Guo nonlinear transformation；
Linear transformation is carried out to the segmentation picture feature Jing Guo dimension-reduction treatment to obtain characteristic color；
Identify that the pedestrian in segmentation picture wears color clothes according to the characteristic color.
10. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that：The computer program
It is executed by processor the method as described in claim 1-5 any one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711059617.0A CN107909580A (en) | 2017-11-01 | 2017-11-01 | A kind of pedestrian wears color identification method, electronic equipment and storage medium clothes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711059617.0A CN107909580A (en) | 2017-11-01 | 2017-11-01 | A kind of pedestrian wears color identification method, electronic equipment and storage medium clothes |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107909580A true CN107909580A (en) | 2018-04-13 |
Family
ID=61842453
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711059617.0A Pending CN107909580A (en) | 2017-11-01 | 2017-11-01 | A kind of pedestrian wears color identification method, electronic equipment and storage medium clothes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107909580A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647689A (en) * | 2018-06-01 | 2018-10-12 | 中国科学院福建物质结构研究所 | Magic square restored method and its device based on GoogLeNet neural networks |
| CN108765449A (en) * | 2018-05-16 | 2018-11-06 | 南京信息工程大学 | A kind of image background segmentation and recognition methods based on convolutional neural networks |
| CN108830908A (en) * | 2018-06-15 | 2018-11-16 | 天津大学 | A kind of magic square color identification method based on artificial neural network |
| CN109273071A (en) * | 2018-08-23 | 2019-01-25 | 广东数相智能科技有限公司 | A method of establishing white object contrast model |
| CN110046577A (en) * | 2019-04-17 | 2019-07-23 | 北京迈格威科技有限公司 | Pedestrian's attribute forecast method, apparatus, computer equipment and storage medium |
| CN110287782A (en) * | 2019-05-17 | 2019-09-27 | 北京迈格威科技有限公司 | Pedestrian's parted pattern training method and device |
| CN110298893A (en) * | 2018-05-14 | 2019-10-01 | 桂林远望智能通信科技有限公司 | A kind of pedestrian wears the generation method and device of color identification model clothes |
| CN110427808A (en) * | 2019-06-21 | 2019-11-08 | 武汉倍特威视***有限公司 | Police uniform recognition methods based on video stream data |
| CN111079757A (en) * | 2018-10-19 | 2020-04-28 | 北京奇虎科技有限公司 | Clothing attribute identification method and device and electronic equipment |
| CN111178128A (en) * | 2019-11-22 | 2020-05-19 | 北京迈格威科技有限公司 | Image recognition method and device, computer equipment and storage medium |
| CN111382746A (en) * | 2018-12-29 | 2020-07-07 | 杭州光启人工智能研究院 | Data marking method, computer device and computer readable storage medium |
| CN111814513A (en) * | 2019-04-11 | 2020-10-23 | 富士通株式会社 | Pedestrian article detection device and method and electronic equipment |
| CN111833340A (en) * | 2020-07-21 | 2020-10-27 | 北京百度网讯科技有限公司 | Image detection method, image detection device, electronic equipment and storage medium |
| CN112232173A (en) * | 2020-10-12 | 2021-01-15 | 济南博观智能科技有限公司 | Pedestrian attribute identification method, deep learning model, equipment and medium |
| CN112489143A (en) * | 2020-11-30 | 2021-03-12 | 济南博观智能科技有限公司 | Color identification method, device, equipment and storage medium |
| CN112560767A (en) * | 2020-12-24 | 2021-03-26 | 南方电网深圳数字电网研究院有限公司 | Document signature identification method and device and computer readable storage medium |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079109A (en) * | 2007-06-26 | 2007-11-28 | 北京中星微电子有限公司 | Identity identification method and system based on uniform characteristic |
| CN104345886A (en) * | 2014-10-09 | 2015-02-11 | 深圳码隆科技有限公司 | Intelligent glasses system for fashion experience and personalized fashion experience method |
| CN105069466A (en) * | 2015-07-24 | 2015-11-18 | 成都市高博汇科信息科技有限公司 | Pedestrian clothing color identification method based on digital image processing |
| CN105224775A (en) * | 2015-11-12 | 2016-01-06 | 中国科学院重庆绿色智能技术研究院 | Based on the method and apparatus that picture processing is arranged in pairs or groups to clothes |
| 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 |
| CN105488490A (en) * | 2015-12-23 | 2016-04-13 | 天津天地伟业数码科技有限公司 | Judge dressing detection method based on video |
| CN105808732A (en) * | 2016-03-10 | 2016-07-27 | 北京大学 | Integration target attribute identification and precise retrieval method based on depth measurement learning |
| CN106250874A (en) * | 2016-08-16 | 2016-12-21 | 东方网力科技股份有限公司 | A kind of dress ornament and the recognition methods of carry-on articles and device |
| CN106504064A (en) * | 2016-10-25 | 2017-03-15 | 清华大学 | Clothes classification based on depth convolutional neural networks recommends method and system with collocation |
| CN106651973A (en) * | 2016-09-28 | 2017-05-10 | 北京旷视科技有限公司 | Image structuring method and device |
| CN106709449A (en) * | 2016-12-22 | 2017-05-24 | 深圳市深网视界科技有限公司 | Pedestrian re-recognition method and system based on deep learning and reinforcement learning |
| US20170169313A1 (en) * | 2015-12-14 | 2017-06-15 | Samsung Electronics Co., Ltd. | Image processing apparatus and method based on deep learning and neural network learning |
| CN106920243A (en) * | 2017-03-09 | 2017-07-04 | 桂林电子科技大学 | The ceramic material part method for sequence image segmentation of improved full convolutional neural networks |
-
2017
- 2017-11-01 CN CN201711059617.0A patent/CN107909580A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079109A (en) * | 2007-06-26 | 2007-11-28 | 北京中星微电子有限公司 | Identity identification method and system based on uniform characteristic |
| CN104345886A (en) * | 2014-10-09 | 2015-02-11 | 深圳码隆科技有限公司 | Intelligent glasses system for fashion experience and personalized fashion experience method |
| 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 |
| CN105069466A (en) * | 2015-07-24 | 2015-11-18 | 成都市高博汇科信息科技有限公司 | Pedestrian clothing color identification method based on digital image processing |
| CN105224775A (en) * | 2015-11-12 | 2016-01-06 | 中国科学院重庆绿色智能技术研究院 | Based on the method and apparatus that picture processing is arranged in pairs or groups to clothes |
| US20170169313A1 (en) * | 2015-12-14 | 2017-06-15 | Samsung Electronics Co., Ltd. | Image processing apparatus and method based on deep learning and neural network learning |
| CN105488490A (en) * | 2015-12-23 | 2016-04-13 | 天津天地伟业数码科技有限公司 | Judge dressing detection method based on video |
| CN105808732A (en) * | 2016-03-10 | 2016-07-27 | 北京大学 | Integration target attribute identification and precise retrieval method based on depth measurement learning |
| CN106250874A (en) * | 2016-08-16 | 2016-12-21 | 东方网力科技股份有限公司 | A kind of dress ornament and the recognition methods of carry-on articles and device |
| CN106651973A (en) * | 2016-09-28 | 2017-05-10 | 北京旷视科技有限公司 | Image structuring method and device |
| CN106504064A (en) * | 2016-10-25 | 2017-03-15 | 清华大学 | Clothes classification based on depth convolutional neural networks recommends method and system with collocation |
| CN106709449A (en) * | 2016-12-22 | 2017-05-24 | 深圳市深网视界科技有限公司 | Pedestrian re-recognition method and system based on deep learning and reinforcement learning |
| CN106920243A (en) * | 2017-03-09 | 2017-07-04 | 桂林电子科技大学 | The ceramic material part method for sequence image segmentation of improved full convolutional neural networks |
Non-Patent Citations (4)
| Title |
|---|
| NANIK SUCIATI等: "Batik Motif Classification Using Color-Texture-Based Feature Extraction and Backpropagation Neural Network", 《2014 IIAI 3RD INTERNATIONAL CONFERENCE ON ADVANCED APPLIED INFORMATICS》 * |
| SHINTAMI C. HIDAYATI等: "Learning and Recognition of Clothing Genres From Full-Body Images", 《IEEE TRANSACTIONS ON CYBERNETICS 》 * |
| 凌穗谊: "基于共现信息及感知归类的衣物识别与解析", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 章星: "基于卷积神经网络的对象颜色识别", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298893A (en) * | 2018-05-14 | 2019-10-01 | 桂林远望智能通信科技有限公司 | A kind of pedestrian wears the generation method and device of color identification model clothes |
| CN108765449B (en) * | 2018-05-16 | 2022-04-26 | 南京信息工程大学 | Image background segmentation and identification method based on convolutional neural network |
| CN108765449A (en) * | 2018-05-16 | 2018-11-06 | 南京信息工程大学 | A kind of image background segmentation and recognition methods based on convolutional neural networks |
| CN108647689A (en) * | 2018-06-01 | 2018-10-12 | 中国科学院福建物质结构研究所 | Magic square restored method and its device based on GoogLeNet neural networks |
| CN108830908A (en) * | 2018-06-15 | 2018-11-16 | 天津大学 | A kind of magic square color identification method based on artificial neural network |
| CN109273071A (en) * | 2018-08-23 | 2019-01-25 | 广东数相智能科技有限公司 | A method of establishing white object contrast model |
| CN111079757A (en) * | 2018-10-19 | 2020-04-28 | 北京奇虎科技有限公司 | Clothing attribute identification method and device and electronic equipment |
| CN111382746A (en) * | 2018-12-29 | 2020-07-07 | 杭州光启人工智能研究院 | Data marking method, computer device and computer readable storage medium |
| CN111814513B (en) * | 2019-04-11 | 2024-02-13 | 富士通株式会社 | Pedestrian article detection device and method and electronic equipment |
| US11488410B2 (en) | 2019-04-11 | 2022-11-01 | Fujitsu Limited | Pedestrian article detection apparatus and method and electronic device |
| CN111814513A (en) * | 2019-04-11 | 2020-10-23 | 富士通株式会社 | Pedestrian article detection device and method and electronic equipment |
| CN110046577A (en) * | 2019-04-17 | 2019-07-23 | 北京迈格威科技有限公司 | Pedestrian's attribute forecast method, apparatus, computer equipment and storage medium |
| CN110287782A (en) * | 2019-05-17 | 2019-09-27 | 北京迈格威科技有限公司 | Pedestrian's parted pattern training method and device |
| CN110427808A (en) * | 2019-06-21 | 2019-11-08 | 武汉倍特威视***有限公司 | Police uniform recognition methods based on video stream data |
| CN111178128A (en) * | 2019-11-22 | 2020-05-19 | 北京迈格威科技有限公司 | Image recognition method and device, computer equipment and storage medium |
| CN111178128B (en) * | 2019-11-22 | 2024-03-19 | 北京迈格威科技有限公司 | Image recognition method, device, computer equipment and storage medium |
| CN111833340A (en) * | 2020-07-21 | 2020-10-27 | 北京百度网讯科技有限公司 | Image detection method, image detection device, electronic equipment and storage medium |
| US11798193B2 (en) | 2020-07-21 | 2023-10-24 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Multi-dimensional image detection on at least two acquired images |
| CN111833340B (en) * | 2020-07-21 | 2024-03-26 | 阿波罗智能技术(北京)有限公司 | Image detection method, device, electronic equipment and storage medium |
| CN112232173A (en) * | 2020-10-12 | 2021-01-15 | 济南博观智能科技有限公司 | Pedestrian attribute identification method, deep learning model, equipment and medium |
| CN112232173B (en) * | 2020-10-12 | 2023-04-07 | 济南博观智能科技有限公司 | Pedestrian attribute identification method, deep learning model, equipment and medium |
| CN112489143A (en) * | 2020-11-30 | 2021-03-12 | 济南博观智能科技有限公司 | Color identification method, device, equipment and storage medium |
| CN112560767A (en) * | 2020-12-24 | 2021-03-26 | 南方电网深圳数字电网研究院有限公司 | Document signature identification method and device and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107909580A (en) | A kind of pedestrian wears color identification method, electronic equipment and storage medium clothes | |
| CN108108761B (en) | Rapid traffic signal lamp detection method based on deep feature learning | |
| CN108765443B (en) | Sign enhancement processing method for self-adaptive color threshold segmentation | |
| CN108549895A (en) | A kind of semi-supervised semantic segmentation method based on confrontation network | |
| CN105354599B (en) | A kind of color identification method based on improved SLIC super-pixel segmentation algorithm | |
| CN105719247B (en) | Single image to the fog method based on feature learning | |
| Sheikh et al. | Traffic sign detection and classification using colour feature and neural network | |
| CN103996041B (en) | Vehicle color identification method and system based on matching | |
| CN109636825A (en) | Seal graphics dividing method, device and computer readable storage medium | |
| CN104899586B (en) | Method and device is identified to the word content for including in image | |
| CN102819728A (en) | Traffic sign detection method based on classification template matching | |
| CN104484658A (en) | Face gender recognition method and device based on multi-channel convolution neural network | |
| Le et al. | Real time traffic sign detection using color and shape-based features | |
| Niu et al. | Image segmentation algorithm for disease detection of wheat leaves | |
| CN105678218B (en) | A kind of method of mobile object classification | |
| Chourasia et al. | Centroid based detection algorithm for hybrid traffic sign recognition system | |
| CN103226696B (en) | The identification system and method for car plate | |
| CN109460715A (en) | A kind of traffic lights automatic identification implementation method based on machine learning | |
| Sanchez-Cuevas et al. | A comparison of color models for color face segmentation | |
| CN107330360A (en) | A kind of pedestrian's clothing colour recognition, pedestrian retrieval method and device | |
| CN110503103A (en) | A kind of character cutting method in line of text based on full convolutional neural networks | |
| CN110298893A (en) | A kind of pedestrian wears the generation method and device of color identification model clothes | |
| CN113283405A (en) | Mask detection method and device, computer equipment and storage medium | |
| CN109284759A (en) | One kind being based on the magic square color identification method of support vector machines (svm) | |
| Lionnie et al. | A comparison of human skin color detection for biometrie identification |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180413 |
|
| RJ01 | Rejection of invention patent application after publication |