CN109035175A - Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network - Google Patents
Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network Download PDFInfo
- Publication number
- CN109035175A CN109035175A CN201810960760.5A CN201810960760A CN109035175A CN 109035175 A CN109035175 A CN 109035175A CN 201810960760 A CN201810960760 A CN 201810960760A CN 109035175 A CN109035175 A CN 109035175A
- Authority
- CN
- China
- Prior art keywords
- color
- facial image
- space
- image
- color correction
- 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
- 230000001815 facial effect Effects 0.000 title claims abstract description 60
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 24
- 238000000034 method Methods 0.000 title claims abstract description 21
- 230000009466 transformation Effects 0.000 claims abstract description 11
- 230000002708 enhancing effect Effects 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 4
- 239000004744 fabric Substances 0.000 description 4
- 238000005286 illumination Methods 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 1
- 241000288673 Chiroptera Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G06T5/92—
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Abstract
The facial image Enhancement Method based on color correction and Pulse Coupled Neural Network that the invention discloses a kind of, the method includes the steps of: carrying out color correction to facial image, facial image partially dark, partially bright, color distortion in image is subjected to color correction, obtains the facial image in the rgb color space being similar under natural scene；Facial image after color correction is transformed into HSI color space from rgb space, obtains tone H, color saturation S and tri- picture contents of brightness I；To luminance component I, it is enhanced using Pulse Coupled Neural Network, obtains enhanced luminance component EnI；Enhanced luminance component EnI and chrominance component H, color saturation component S are done into the inverse transformation of HIS space to rgb space, obtain enhanced facial image in rgb space.The beneficial effects of the present invention are image color difference problem is effectively improved, enhance the details of facial image.
Description
Technical field
The present invention relates to a kind of methods of facial image enhancing, and in particular to one kind is based on color correction and pulse-couple mind
Facial image Enhancement Method through network.
Background technique
Facial image enhancing is a research branch of field of image enhancement, is that present mode identification is ground with what information identified
Study carefully one of hot spot.In numerous biological information authentication techniques, recognition of face is a kind of technology with numerous advantages, face
Identification plays an important role in the application such as authentication, human-computer interaction.But numerous face recognition technologies are to facial image
It is more demanding, such as illumination, facial angle, face image quality etc., therefore, it is still one important that facial image, which enhances technology,
Hot research field.
Facial image enhancing is usually to handle the facial image object in video, such as the people in dynamic video sequence
Facial image under face image or static scene.The purpose of facial image enhancing is made in different illumination, different angle, different bats
The facial image under environment is taken the photograph by color correction, carries out image detail enhancing, is into one so that picture quality be made to get a promotion
Recognition of face, face classification of step etc. carry out element task.It is usually direct to facial image in the application of actual image enhancement
Enhanced, some problems can be brought in this way, such as collected facial image increases in the case where night or bad illumination condition
Noise after strong in same meeting enlarged drawing, can not be effectively improved image definition.In recent years, it proposes many in different color sky
Between under image enchancing method, such as under HSI, HSV color space, the luminance component of image is enhanced, these enhancings
Method can be effectively improved the clarity of image, but for there is the facial image of color difference that can not be effectively improved the color point of image
Cloth situation.Other image enchancing methods based on Multiresolution Decomposition, this kind of algorithm decompose original image in multiscale space,
The image detail of different levels is enhanced respectively, such algorithm can also effectively enhance image detail information, but for color
Situations such as coloured silk distortion, is still unable to improve.
Summary of the invention
Present invention is generally directed to facial images to enhance problem, and to solve the above problems existing in the prior art, the present invention mentions
A kind of facial image Enhancement Method based on color correction and Pulse Coupled Neural Network is supplied.
To achieve the goals above, the invention adopts the following technical scheme:
Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network, the present invention the following steps are included:
Step 1 carries out color correction to facial image, and facial image partially dark, partially bright, color distortion in image is carried out
Color correction obtains the facial image in the rgb color space being similar under natural scene.
Step 1 comprises the steps of:
Step 1.1, the face reference image R under a natural scene is chosen
Step 1.2, by facial image I to be reinforced and reference image R, formula (1)-(3) are sequentially utilized:
Be transformed into l α β color space from rgb color space, respectively obtain three component l, α in l α β color space,
β, wherein l indicates brightness (Luminosity), and α indicates the range from carmetta to green, and β indicates the model from yellow to blue
It encloses.
Step 1.3, calculate separately image I to be reinforced and reference image R in l α β color space the mean value of three components with
Variance, and use respectivelyWithIt indicates.
Step 1.4, using formula (4), according to the COLOR COMPOSITION THROUGH DISTRIBUTION mode of reference image R, to the color point of image I to be reinforced
Cloth is converted, and obtains being similar to the facial image of COLOR COMPOSITION THROUGH DISTRIBUTION under natural scene:
Wherein, l*、α*、β*Three components in l α β color space after respectively indicating conversion.
Step 1.5, formula (5)-(7) are sequentially utilized, rgb space is changed in the face figure contravariant after COLOR COMPOSITION THROUGH DISTRIBUTION is corrected,
Face figure after obtaining color correction:
Facial image after color correction is transformed to HSI color space from rgb space by step 2, obtains H, S, I tri- figures
As component.
Step 3 enhances it to luminance component I, using Pulse Coupled Neural Network, obtains enhanced brightness point
Measure EnI.
Step 4 by enhanced luminance component EnI and step 2 chrominance component H and color saturation component S carry out HSI
Space obtains enhanced facial image in rgb space to the inverse transformation of rgb space.
Using the device based on color correction and the facial image Enhancement Method of Pulse Coupled Neural Network, which includes
Following hardware module；1) the color correction module of step 1 function is realized；2) the colour space transformation mould of step 2 function is realized
Block；3) realize that the luminance component of step 3 function enhances module；4) the color space inverse transform module of step 4 function is realized.
The connection relationship of the above-mentioned device of the present invention is that color correction module, colour space transformation module, luminance component increase
Strong module, color space inverse transform module sequentially connect.
The beneficial effects of the present invention are: propose a kind of face figure based on color correction and Pulse Coupled Neural Network
Image intensifying algorithm, specifically:
(1) facial image is subjected to color correction first, specially original image is transformed into l α β color space, according to ginseng
The COLOR COMPOSITION THROUGH DISTRIBUTION situation for examining image (standard faces image) carries out color correction to original image, obtains being similar to reference picture color point
The facial image of cloth.The problems such as this step can be effectively improved image color distortion, color difference is obvious.
(2) image after color correction is transformed into the space HSI, Pulse Coupled Neural Network is utilized to luminance component I
Enhanced.Luminance component I is enhanced, image color distortion bring can not can effectively be overcome to ask to avoid traditional algorithm
Topic.
(3) the chrominance component H of the image after color correction, color saturation component S and enhanced luminance component I are carried out
HSI inverse transformation is transformed into rgb color space, the facial image finally enhanced.
The problems such as being handled by images above enhancing, the color distortion and color difference of facial image can be effectively improved, enhancing
The details of facial image lays the foundation for technologies such as recognitions of face.
The present invention will be further described with reference to the accompanying drawings and detailed description.
Detailed description of the invention
Fig. 1 is that the present inventor's face image enhances algorithm flow chart.
Fig. 2 is that original image is carried out color correction and enhanced comparison diagram using method of the invention.
Specific embodiment
Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network, comprising the following steps: to original
The problems such as facial image carries out color correction, improves image color distortion, color difference；Image after color correction is transformed into HSI
In space, three components are obtained；To luminance component, enhanced by the way of Pulse Coupled Neural Network；It will be enhanced
H and S component after luminance component and color correction carry out inverse transformation, are transformed into rgb space, obtain enhancing facial image.Specifically
Are as follows:
(1) color correction is carried out to facial image, steps are as follows for specific color correction:
Step 1.1, the face reference image R under a natural scene is chosen
Step 1.2, by facial image I to be reinforced and reference image R, formula (1)-(3) are sequentially utilized:
Be transformed into l α β color space from rgb color space, respectively obtain three component l, α in l α β color space,
β.Step 1.3, the mean value and variance of image I to be reinforced and reference image R three components in l α β color space are calculated separately,
And it uses respectivelyWithIt indicates.
Step 1.4, using formula (4), according to the COLOR COMPOSITION THROUGH DISTRIBUTION mode of reference image R, to the color point of image I to be reinforced
Cloth is converted, and obtains being similar to the facial image of COLOR COMPOSITION THROUGH DISTRIBUTION under natural scene:
Step 1.5, formula (5)-(7) are sequentially utilized, rgb space is changed in the face figure contravariant after COLOR COMPOSITION THROUGH DISTRIBUTION is corrected,
Face figure after obtaining color correction.
(2) facial image is decomposed in the space HSI, obtains tri- picture contents of H, S, I；
(3) luminance component I is enhanced using following formula:
EnIij=ln (Bri)-(k-1) aθ
In formula, EnIijFor enhanced luminance component, Bri is the max pixel value in image I to be reinforced, and k is pulse coupling
Close the number of iterations when neural network igniting, aθIt is Pulse Coupled Neural Network parameter；
(4) by enhanced luminance component EnIij, inverse transformation is carried out with chrominance component H and color saturation component S, is transformed into
In rgb space, enhanced facial image is obtained.
Using the device based on color correction and the facial image Enhancement Method of Pulse Coupled Neural Network, which includes
Following hardware module；1) the color correction module of step 1 function is realized；2) the colour space transformation mould of step 2 function is realized
Block；3) realize that the luminance component of step 3 function enhances module；4) the color space inverse transform module of step 4 function is realized.
The connection relationship of the above-mentioned device of the present invention is that color correction module, colour space transformation module, luminance component increase
Strong module, color space inverse transform module sequentially connect.
Letter refers to explanation in the present invention:
1.I: facial image (Image initial) to be reinforced；
2.R: reference picture (Reference initial)；
The space 3.RGB: refer to the color model based on tri- kinds of Essential colour of R, G, B.R indicates red (Red) that G indicates green
Color (Green), B indicate blue (Blue)；R, G, B are color component.
The space 4.HSI: refer to tri- kinds of essential characteristic amounts of H, S and I and perceive the model of color.H indicates chrominance component
(Hue), S indicates color saturation component (Saturation), and I indicates luminance component (Intensity).
The space 5.l α β: referring to by brightness (l) and tri- elements of α, β in relation to color form.L indicates brightness
(Luminosity), α indicates the range from carmetta to green, and β indicates the range from yellow to blue.
6.L, M, S: being 3 intermediate quantities, indicates to carry out edited point of non-linear tone to three components in rgb space
Amount.
L', M', S': being the component of intermediate quantity L, M, S after logarithmic transformation.
7.Respectively indicate the mean value of image I to be reinforced three components l, α, β in l α β color space；
Respectively indicate the variance of image I to be reinforced three components l, α, β in l α β color space；
Respectively indicate the mean value of reference image R three components l, α, β in l α β color space；
Respectively indicate the variance of reference image R three components l, α, β in l α β color space.
8.l*、α*、β*: it indicates the COLOR COMPOSITION THROUGH DISTRIBUTION mode according to reference image R, the COLOR COMPOSITION THROUGH DISTRIBUTION of image I to be reinforced is carried out
Three components in l α β color space after conversion.
9.EnI: enhanced luminance component；
EnIij: the element that the i-th row jth arranges in enhanced luminance component EnI.
10.Bri: the max pixel value in image I to be reinforced
11.k: the number of iterations when Pulse Coupled Neural Network is lighted a fire.
12.aθ: Pulse Coupled Neural Network parameter.
Claims (5)
1. the facial image Enhancement Method based on color correction and Pulse Coupled Neural Network, which is characterized in that including following step
It is rapid:
Step 1 carries out color correction to facial image, and facial image partially dark, partially bright, color distortion in image is carried out color
It corrects, obtains the facial image in the rgb color space being similar under natural scene；
Facial image after color correction is transformed to HSI color space from rgb space by step 2, obtains tri- images of H, S, I point
Amount；
Step 3 enhances it to luminance component I, using Pulse Coupled Neural Network, obtains enhanced luminance component
EnI；
Step 4 by enhanced luminance component EnI and step 2 chrominance component H and color saturation component S carry out the space HSI
To the inverse transformation of rgb space, enhanced facial image in rgb space is obtained.
2. the facial image Enhancement Method according to claim 1 based on color correction and Pulse Coupled Neural Network,
It is characterized in that, step 1 comprises the steps of:
Step 1.1, the face reference image R under a natural scene is chosen
Step 1.2, by facial image I to be reinforced and reference image R, formula (1)-(3) are sequentially utilized:
It is transformed into l α β color space from rgb color space, respectively obtains three components l, α, β in l α β color space,
In, l indicates that brightness Luminosity, α indicate the range from carmetta to green, and β indicates the range from yellow to blue；Step
1.3, the mean value and variance of image I to be reinforced and reference image R three components in l α β color space are calculated separately, and respectively
WithWithIt indicates；
Step 1.4, using formula (4), according to the COLOR COMPOSITION THROUGH DISTRIBUTION mode of reference image R, to the COLOR COMPOSITION THROUGH DISTRIBUTION of image I to be reinforced into
Row conversion, obtains being similar to the facial image of COLOR COMPOSITION THROUGH DISTRIBUTION under natural scene:
Wherein, l*、α*、β*Three components in l α β color space after respectively indicating conversion；
Step 1.5, formula (5)-(7) are sequentially utilized, the face figure contravariant after COLOR COMPOSITION THROUGH DISTRIBUTION is corrected is changed to rgb space, obtained
Face figure after color correction:
3. the facial image Enhancement Method according to claim 1 based on color correction and Pulse Coupled Neural Network,
It is characterized in that, step 3 is the following steps are included: enhance luminance component I using following formula:
EnIij=ln (Bri)-(k-1) aθ
In formula, EnIijFor the element that the i-th row jth in enhanced luminance component arranges, Bri is the maximum picture in image I to be reinforced
Element value, the number of iterations when k is Pulse Coupled Neural Network igniting, aθIt is Pulse Coupled Neural Network parameter.
4. using the device of the facial image Enhancement Method based on color correction and Pulse Coupled Neural Network, which is characterized in that
The device includes following hardware module；1) the color correction module of step 1 function is realized；2) color of step 2 function is realized
Spatial alternation module；3) realize that the luminance component of step 3 function enhances module；4) realize that the color space of step 4 function is inverse
Conversion module.
5. device according to claim 4, which is characterized in that the connection relationship of the device is color correction module, color
Spatial alternation module, luminance component enhancing module, color space inverse transform module sequentially connect.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810960760.5A CN109035175A (en) | 2018-08-22 | 2018-08-22 | Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810960760.5A CN109035175A (en) | 2018-08-22 | 2018-08-22 | Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109035175A true CN109035175A (en) | 2018-12-18 |
Family
ID=64626950
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810960760.5A Pending CN109035175A (en) | 2018-08-22 | 2018-08-22 | Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109035175A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109669660A (en) * | 2019-02-28 | 2019-04-23 | 深圳市华星光电半导体显示技术有限公司 | Image processing system and image processing method |
| CN112069475A (en) * | 2020-09-14 | 2020-12-11 | 杭州熙菱信息技术有限公司 | Identity safety management system |
| WO2022062346A1 (en) * | 2020-09-27 | 2022-03-31 | 苏州科达科技股份有限公司 | Image enhancement method and apparatus, and electronic device |
| WO2024027583A1 (en) * | 2022-08-03 | 2024-02-08 | 维沃移动通信有限公司 | Image processing method and apparatus, and electronic device and readable storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030053690A1 (en) * | 2001-07-06 | 2003-03-20 | Jasc Software, Inc. | Automatic contrast enhancement |
| CN102779330A (en) * | 2012-06-13 | 2012-11-14 | 京东方科技集团股份有限公司 | Image reinforcement method, image reinforcement device and display device |
| CN103345732A (en) * | 2013-07-26 | 2013-10-09 | 电子科技大学 | Pulse coupled neural network (PCNN) image enhancement algorithm and device based on Contourlet transformation |
| CN103632339A (en) * | 2012-08-21 | 2014-03-12 | 张晓光 | Single image defogging method based on variation Retinex and apparatus |
-
2018
- 2018-08-22 CN CN201810960760.5A patent/CN109035175A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030053690A1 (en) * | 2001-07-06 | 2003-03-20 | Jasc Software, Inc. | Automatic contrast enhancement |
| CN102779330A (en) * | 2012-06-13 | 2012-11-14 | 京东方科技集团股份有限公司 | Image reinforcement method, image reinforcement device and display device |
| CN103632339A (en) * | 2012-08-21 | 2014-03-12 | 张晓光 | Single image defogging method based on variation Retinex and apparatus |
| CN103345732A (en) * | 2013-07-26 | 2013-10-09 | 电子科技大学 | Pulse coupled neural network (PCNN) image enhancement algorithm and device based on Contourlet transformation |
Non-Patent Citations (2)
| Title |
|---|
| 钱文华等: ""lαβ空间中纹理传递的实现"", 《第一届建立和谐人机环境联合学术会议（HHME2005）论文集》 * |
| 马振兴: ""基于脉冲耦合神经网络的医学图像增强处理"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109669660A (en) * | 2019-02-28 | 2019-04-23 | 深圳市华星光电半导体显示技术有限公司 | Image processing system and image processing method |
| CN112069475A (en) * | 2020-09-14 | 2020-12-11 | 杭州熙菱信息技术有限公司 | Identity safety management system |
| CN112069475B (en) * | 2020-09-14 | 2023-10-24 | 杭州领信数科信息技术有限公司 | Identity security management system |
| WO2022062346A1 (en) * | 2020-09-27 | 2022-03-31 | 苏州科达科技股份有限公司 | Image enhancement method and apparatus, and electronic device |
| WO2024027583A1 (en) * | 2022-08-03 | 2024-02-08 | 维沃移动通信有限公司 | Image processing method and apparatus, and electronic device and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110047051B (en) | Non-uniform illumination color image enhancement method | |
| CN109035175A (en) | Facial image Enhancement Method based on color correction and Pulse Coupled Neural Network | |
| CN106504212A (en) | A kind of improved HSI spatial informations low-luminance color algorithm for image enhancement | |
| CN103593830B (en) | A kind of low illumination level video image enhancement | |
| CN107038680B (en) | Self-adaptive illumination beautifying method and system | |
| CN100412906C (en) | Method for correcting digital tongue picture colour cast | |
| CN106897981A (en) | A kind of enhancement method of low-illumination image based on guiding filtering | |
| CN106530250A (en) | Low illumination color image enhancement method based on improved Retinex | |
| US7933469B2 (en) | Video processing | |
| WO2004098167A3 (en) | A system and process for generating high dynamic range video | |
| WO2013185449A1 (en) | Image enhancement method, image enhancement device and display device | |
| CN101930596A (en) | Color constancy method in two steps under a kind of complex illumination | |
| EP1918871A3 (en) | Image enhancement method and system | |
| CN108288258A (en) | A kind of low-quality images Enhancement Method under severe weather conditions | |
| CN106997584A (en) | A kind of haze weather image enchancing method | |
| CN111105359B (en) | Tone mapping method for high dynamic range image | |
| US20080056566A1 (en) | Video processing | |
| CN109218716A (en) | Based on color statistics and comentropy without reference tone mapping graph image quality evaluation method | |
| CN107392879B (en) | A kind of low-light (level) monitoring image Enhancement Method based on reference frame | |
| CN104778674B (en) | A kind of suitable backlight traffic image self-adapting enhancement method based on time series | |
| CN106340025A (en) | Background replacement visual communication method based on chromatic adaptation transformation | |
| CN106485674A (en) | A kind of low light image Enhancement Method based on integration technology | |
| CN109658367A (en) | Image interfusion method based on Color transfer | |
| CN104112259A (en) | Rain removing method and system for single image | |
| CN109636739B (en) | Detail processing method and device for enhancing image saturation |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181218 |