CN106469300B - A kind of color spot detection recognition method - Google Patents
A kind of color spot detection recognition method Download PDFInfo
- Publication number
- CN106469300B CN106469300B CN201610798036.8A CN201610798036A CN106469300B CN 106469300 B CN106469300 B CN 106469300B CN 201610798036 A CN201610798036 A CN 201610798036A CN 106469300 B CN106469300 B CN 106469300B
- Authority
- CN
- China
- Prior art keywords
- color
- color spot
- average
- spot
- information
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of color spot detection recognition method, gathers face picture, and face picture includes the mutation picture of all color spot species and color spot；And color spot region and color spot species are marked in advance, by being read out and being iterated to the feature vector one and markup information of each color spot until obtaining the color characteristic average and Wavelet Transform Feature average of each color spot, the feature vector two and markup information of each color spot are read out and are iterated until obtaining color average, texture average and the blob features average of each color spot, its iterative process uses SVM；The color spot information in face picture to be detected is finally read, carries out matching acquisition color spot region with the color characteristic average and Wavelet Transform Feature average；Match cognization color spot species is carried out with color average, texture average and blob features average；Export the color spot region obtained and color spot kind of information.Color spot detection recognition method provided by the invention has good migration, and precision is high.
Description
Technical field
The present invention relates to beautifying medical field, particularly relates to a kind of color spot detection recognition method.
Background technology
The technology that traditional color spot detection and recognition methods are split or strengthened using image more obtains, and binding rule carries out color
Spot detects, and in terms of identification, obtains the species of color spot using template matches either neighborhood matching more.Above method requires image
There is good severability, but in reality scene, due to the influence of many factors such as capture apparatus, illumination, to image into acting charitably
Segmentation it is relatively difficult, cause color spot detection and identification error it is larger；Simultaneously as partitioning algorithm needs manually to carry out parameter tune
It is whole, and parameter majority is manually adjusted by being carried out on a certain number of sampled images, it is impossible to reflect the distribution of image entirety, institute
It is poor with the transportable property of partitioning algorithm.
The content of the invention
The present invention proposes that one kind has good migration, and the color spot detection recognition method that precision is high.
The technical proposal of the invention is realized in this way：
A kind of color spot detection recognition method, comprises the following steps：
Step 1：N face pictures, wherein N >=1000 are gathered, face picture includes the change of all color spot species and color spot
Kind picture；
Step 2：Color spot region and color spot are marked according to mode set in advance to the color spot in each face picture
Species, sets the information as markup information and is preserved；
Step 3：According to the color spot detection model of pixel, the color characteristic and Wavelet Transform Feature of color spot are extracted, is set
The information is preserved in the lump for feature vector；
Step 4：The feature vector one and markup information of each color spot are read out and are iterated until obtaining each
The color characteristic average and Wavelet Transform Feature average of kind color spot are simultaneously preserved, its iterative process uses SVM (supporting vectors
Machine)；
Step 5：Each color spot point centered on its center, and to surround the minimum rectangle of color spot extraction color spot picture,
The unified deformation (warping) of these pictures is arrived to the square of same size；Color spot figure after extraction deformation (warping)
Color, texture and blob features in piece simultaneously set the category information as feature vector two；
Step 6：The feature vector two and markup information of each color spot are read out and are iterated until obtaining each
Color average, texture average and the blob features average of kind color spot are simultaneously preserved, its iterative process uses SVM (supporting vectors
Machine)；
Step 7：The color spot information in face picture to be detected is read, morphologic denoising, bonding are carried out to it, is put down
With the color characteristic average and Wavelet Transform Feature average match after sliding processing and obtain color spot region；It is again that this is to be detected
Color spot information and color average, texture average and blob features average carry out match cognization color spot species；Export the color obtained
Spot region and color spot kind of information.
A kind of color spot detection recognition method provided by the invention, face picture database i.e. color is established by step 1 and two
Spot database, by step 3 and one information of feature vector in four analysis color spot databases, obtains color spot regional model, passes through
Step 5 and two information of feature vector in step 6 analysis color spot database, obtain color spot kind class model, then according to region
The face picture that model and color spot kind class model treat detection is identified, so as to complete high-precision detection and identification,
The present invention by training pattern by applying to color spot detection field, while by the very effective color of computer vision field, line
The parameter that the features such as reason, small echo are considered as identification, so as to improve the precision of identification；And this method can also be further
Apply to color spot latent structure and combinations of features direction, therefore there is good migration.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is attached drawing needed in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is a kind of flow chart of color spot detection recognition method of the present invention.
Embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without creative efforts
Embodiment, belongs to the scope of protection of the invention.
Fig. 1 is referred to, a kind of color spot detection recognition method, comprises the following steps：
Step 1：N face pictures, wherein N >=1000 are gathered, face picture includes the change of all color spot species and color spot
Kind picture；Bigger in the numerical value of the various N of this step, species is then more complete, and can be the essence that follow-up execution step lifts higher
Degree.
Step 2：Color spot region and color spot are marked according to mode set in advance to the color spot in each face picture
Species, sets the information as markup information and is preserved；This step can preset different notation methods.Its notation methods
Need to meet：1st, labeled standards must be unified；2nd, need mark two information, one be color spot region, by irregular
Curve marks, the other is the color spot species in the corresponding color spot region of mark.Annotation results, which preserve, is divided into two parts, color spot area
Domain directly marked out on the diagram come, color spot species and corresponding picture name in excel with cognizable colour curve or
Preserve and correspond in person's txt texts.
Step 3：According to the color spot detection model of pixel, the color characteristic and Wavelet Transform Feature of color spot are extracted, is set
The information is preserved in the lump for feature vector；The color spot detection model based on pixel is judged in this step, that is, is sentenced
Whether some disconnected pixel belongs to color spot region, and judges whether a pixel belongs to color spot region, and the face of itself
The color and vein information of color information, texture information and peripheral cell domain is all related.Therefore extraction color spot and its peripheral cell
The color characteristic and Wavelet Transform Feature in domain.
Step 4：The feature vector one and markup information of each color spot are read out and are iterated until obtaining each
The color characteristic average and Wavelet Transform Feature average of kind color spot are simultaneously preserved, its iterative process uses SVM (supporting vectors
Machine)；This step also referred to as uses SVM (support vector machines) train classification models, each color spot it includes many kinds of parameters, its
Parameter is constantly adjusted during iteration, final to obtain color characteristic average and Wavelet Transform Feature average；The color
Characteristic mean and Wavelet Transform Feature average are the optimum value close to the actual parameter of corresponding color spot, and it is a region model
Enclose value.
Step 5：Each color spot point centered on its center, and to surround the minimum rectangle of color spot extraction color spot picture,
The unified deformation (warping) of these pictures is arrived to the square of same size；Color spot figure after extraction deformation (warping)
Color, texture and blob features in piece simultaneously set the category information as feature vector two；Color spot center is external for its in this step
The center of rectangle.And use minimum rectangle extraction color spot picture be due to it includes color spot Information Meter it is the most accurate.
Step 6：The feature vector two and markup information of each color spot are read out and are iterated until obtaining each
Color average, texture average and the blob features average of kind color spot are simultaneously preserved, its iterative process uses SVM；With step 4
Principle is similar, which also also referred to as uses SVM (support vector machines) train classification models, but its parameter is different, it reads
Numerical value be the numerical value of the color spot picture after deformation (warping), the color average of the color spot of acquisition, texture average and spot are special
It is also the optimum value close to the actual parameter of corresponding color spot to levy average, and it is a regional extent value.
Step 7：The color spot information in face picture to be detected is read, morphologic denoising, bonding are carried out to it, is put down
With the color characteristic average and Wavelet Transform Feature average match after sliding processing and obtain color spot region；It is again that this is to be detected
Color spot information and color average, texture average and blob features average carry out match cognization color spot species；Export the color obtained
Spot region and color spot kind of information.Wherein carry out morphologic denoising, bonding, smoothing processing are a kind of basic operation hand in this area
Section, this step can realize the automatic process detected with identification, and by the very effective color of computer vision field, texture, small
The parameter that the features such as ripple are considered as identification, so as to improve the precision of identification.
Preferably, all color spot species can be divided into active spot, qualitative spot according to property；It can also be included according to classification common
Freckle, blackspot, chloasma etc.；And if when there is new spot kind, can also be increased by way of repeat step one to seven
The characteristic information of the spot kind, so as to fulfill the function of detecting and identify.
Preferably, when performing step 2, mode set in advance is that different face is preset according to different color spots
Color is classified with numbering.Its certain not limited to this mode classification, it can also classify according only to the mode of numbering,
A kind of color spot detection recognition method provided by the invention, face picture database i.e. color is established by step 1 and two
Spot database, its process can be preset, and are labeled operation by professional person and then be can further improve precision, be led to
Step 3 and one information of feature vector in four analysis color spot databases are crossed, obtains color spot regional model, this process can basis
Need appropriate increase number；Two information of feature vector in color spot database is analyzed by step 5 and step 6, obtains color
Spot kind class model, this process can also appropriate increase numbers as needed；Then according to regional model and color spot kind class model
Face picture to be detected is identified, so as to complete it is high-precision detection and identification, the present invention by will training mould
Type applies to color spot detection field, while using features such as the very effective color of computer vision field, texture, small echos as knowledge
The parameter do not considered, so as to improve the precision of identification；And this method can also further apply to color spot feature structure
Make with combinations of features direction, therefore there is good migration, and the present invention has extremely strong applicability in beautifying medical field；
So as to greatly improve the efficiency of doctor, and the precision judged, and the database for establishing by this mode bigger provides
Source, more science, break away from traditional subjective experience for only leaning on experience and domain knowledge.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
With within principle, any modification, equivalent replacement, improvement and so on, should all be included in the protection scope of the present invention god.
Claims (2)
1. a kind of color spot detection recognition method, it is characterised in that comprise the following steps：
Step 1：N face pictures, wherein N >=1000 are gathered, face picture includes the mutation figure of all color spot species and color spot
Piece；
Step 2：Color spot region and color spot kind are marked according to mode set in advance to the color spot in each face picture
Class, sets the information as markup information and is preserved；
Step 3：According to the color spot detection model of pixel, the color characteristic and Wavelet Transform Feature of color spot are extracted, sets the letter
Cease and preserved in the lump for feature vector；
Step 4：The feature vector one and markup information of each color spot are read out and are iterated until obtaining each color
The color characteristic average and Wavelet Transform Feature average of spot are simultaneously preserved, its iterative process uses SVM；
Step 5：Each color spot point centered on its center, and to surround the minimum rectangle of color spot extraction color spot picture, by this
A little pictures are uniformly deformed to the square of same size；Extract color, texture and the spot in deformed color spot picture
Feature simultaneously sets the category information as feature vector two；
Step 6：The feature vector two and markup information of each color spot are read out and are iterated until obtaining each color
Color average, texture average and the blob features average of spot are simultaneously preserved, its iterative process uses SVM；
Step 7：The color spot information in face picture to be detected is read, morphologic denoising, bonding, smooth place are carried out to it
With the color characteristic average and Wavelet Transform Feature average match after reason and obtain color spot region；Again by the color to be detected
Spot information carries out match cognization color spot species with color average, texture average and blob features average；Export the color spot area obtained
Domain and color spot kind of information.
2. a kind of color spot detection recognition method as claimed in claim 1, it is characterised in that when performing step 2, set in advance
Fixed mode is classified to preset different colors according to different color spots with numbering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610798036.8A CN106469300B (en) | 2016-08-31 | 2016-08-31 | A kind of color spot detection recognition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610798036.8A CN106469300B (en) | 2016-08-31 | 2016-08-31 | A kind of color spot detection recognition method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106469300A CN106469300A (en) | 2017-03-01 |
| CN106469300B true CN106469300B (en) | 2018-05-08 |
Family
ID=58230432
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610798036.8A Active CN106469300B (en) | 2016-08-31 | 2016-08-31 | A kind of color spot detection recognition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106469300B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109299634A (en) * | 2017-07-25 | 2019-02-01 | 上海中科顶信医学影像科技有限公司 | Spot detection method, system, equipment and storage medium |
| CN108780497B (en) * | 2017-11-14 | 2021-10-26 | 深圳和而泰智能控制股份有限公司 | Skin flaw point classification method and electronic equipment |
| CN108846311A (en) * | 2018-04-28 | 2018-11-20 | 北京羽医甘蓝信息技术有限公司 | The method and device of the facial pieces of skin shape defect of detection based on deep learning |
| CN109529202B (en) * | 2018-12-29 | 2023-04-07 | 佛山科学技术学院 | Laser speckle removing system and method |
| CN111429416B (en) * | 2020-03-19 | 2023-10-13 | 深圳数联天下智能科技有限公司 | Facial pigment spot recognition method and device and electronic equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102486793A (en) * | 2010-12-06 | 2012-06-06 | 武汉若鱼网络科技有限公司 | Method and system for searching target user |
| CN105869159A (en) * | 2016-03-28 | 2016-08-17 | 联想(北京)有限公司 | Image segmentation method and apparatus |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8218862B2 (en) * | 2008-02-01 | 2012-07-10 | Canfield Scientific, Incorporated | Automatic mask design and registration and feature detection for computer-aided skin analysis |
-
2016
- 2016-08-31 CN CN201610798036.8A patent/CN106469300B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102486793A (en) * | 2010-12-06 | 2012-06-06 | 武汉若鱼网络科技有限公司 | Method and system for searching target user |
| CN105869159A (en) * | 2016-03-28 | 2016-08-17 | 联想(北京)有限公司 | Image segmentation method and apparatus |
Non-Patent Citations (1)
| Title |
|---|
| 一种基于人脸皮肤图像的色斑检测算法;储霞等;《微计算机信息》;20090731;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106469300A (en) | 2017-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106469300B (en) | A kind of color spot detection recognition method | |
| CN109284758B (en) | Invoice seal eliminating method and device and computer storage medium | |
| CN103839223B (en) | Image processing method and device | |
| CN105447441B (en) | Face authentication method and device | |
| CN109410026A (en) | Identity identifying method, device, equipment and storage medium based on recognition of face | |
| CN107871125A (en) | Architecture against regulations recognition methods, device and electronic equipment | |
| CN104036278B (en) | The extracting method of face algorithm standard rules face image | |
| CN107844736B (en) | Iris positioning method and device | |
| CN100382751C (en) | Canthus and pupil location method based on VPP and improved SUSAN | |
| CN108615239B (en) | Tongue image segmentation method based on threshold technology and gray level projection | |
| CN110097034A (en) | A kind of identification and appraisal procedure of Intelligent human-face health degree | |
| TWI709085B (en) | Method, device, computer readable storage medium and computing equipment for damage segmentation of vehicle damage image | |
| CN103824059A (en) | Facial expression recognition method based on video image sequence | |
| CN106446952A (en) | Method and apparatus for recognizing score image | |
| CN103218605A (en) | Quick eye locating method based on integral projection and edge detection | |
| CN108846311A (en) | The method and device of the facial pieces of skin shape defect of detection based on deep learning | |
| CN109086772A (en) | A kind of recognition methods and system distorting adhesion character picture validation code | |
| CN104794693A (en) | Human image optimization method capable of automatically detecting mask in human face key areas | |
| CN108875645A (en) | A kind of face identification method under the conditions of underground coal mine complex illumination | |
| Vishwakarma et al. | Simple and intelligent system to recognize the expression of speech-disabled person | |
| CN110472605A (en) | A kind of skin problem diagnostic method based on deep learning face subregion | |
| CN107169427A (en) | One kind is applied to psychologic face recognition method and device | |
| CN109410138A (en) | Modify jowled methods, devices and systems | |
| CN110298835A (en) | A kind of leather surface damage testing method, system and relevant apparatus | |
| CN110956184A (en) | Abstract diagram direction determination method based on HSI-LBP characteristics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |