CN104732214A - Quantification skin detecting method based on face image recognition - Google Patents
Quantification skin detecting method based on face image recognition Download PDFInfo
- Publication number
- CN104732214A CN104732214A CN201510130557.1A CN201510130557A CN104732214A CN 104732214 A CN104732214 A CN 104732214A CN 201510130557 A CN201510130557 A CN 201510130557A CN 104732214 A CN104732214 A CN 104732214A
- Authority
- CN
- China
- Prior art keywords
- face
- image block
- skin
- point
- score value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
The invention discloses a quantification skin detecting method based on face image recognition. The quantification skin detecting method includes the steps that two clear face pictures are obtained through photographing, wherein one picture is obtained when a flash lamp is started, and the other picture is obtained under the good illumination condition when the flash lamp is not started; face recognition is carried out on the face pictures, and each face obtained after recognition is divided into eight face image blocks; skin attribute calculation is carried out on the face image blocks generated after dividing, wherein the skin attributes include the glossy degree score, the skin smooth degree score, the skin pore coarse degree score and the total blackhead serious degree score; the skin type is detected and judged according to the results of the skin attributes of the blocks. By means of the quantification skin detecting method, the face pictures can be analyzed, the skin of a person can be detected through the face pictures, and a skin caring method and skin caring suggestions can be provided for the next step; the quantification skin detecting method can be widely applied to the skin caring field, and people can be conveniently and rapidly assisted to make correct skin caring choices at any time and any place through recognition and detection of the face pictures.
Description
Technical field
The present invention relates to a kind of recognition detection method of facial image, especially related to a kind of quantification skin quality detection method based on facial image identification, belong to the field of image recognition of artificial intelligence.
Background technology
Recognition of face is a kind of biological identification technology carrying out identification based on the face feature information of people.Image or the video flowing of face is contained with video camera or camera collection, and automatic detection and tracking face in the picture, and then the face detected is carried out to a series of correlation techniques of face, be usually also called Identification of Images, face recognition.Current face recognition technology comparative maturity, but the skin skin quality that also cannot be detected a people by the face in image quantitatively.Skin quality refers to the specific properties that the variation of human skin is formed and feature.Main difficulty has:
(1) the skin quality situation of people is complicated, and the situation of skin varies with each individual.This species diversity is mainly caused by the vigorous degree of the smegma sebum under epidermal area.(main chemical compositions of sebum is triglyceride, and other also have wax ester, MF59 and cholesteryl ester etc., is also commonly called as skin oil, fuel-displaced.) according to facial epidermis different parts sebum secretion situation, neutral skin quality, dry skin, oiliness skin quality, Combination skin quality can be divided into.Wherein combination is at facial T district (volume, nose, mouth, lower jaw) in oiliness, and all the other positions are dryness.Also lacking special image-recognizing method can identify this four kinds of skin quality at present.
(2) even same skin quality type, due to the vigorous degree difference of epidermis smegma of individuality, and the difference of personal nursing skin, also may show different skin condition.Such as, be all Combination skin quality, somebody's chin sebum secreted is vigorous, and somebody is then less in chin place sebum secreted.Another one example is, is all oiliness skin quality, and it is thick that the people being negligent of nursing usually there will be pore, and the situation of long blackhead, good people then must not have this problem to skin nursing.If accomplish to provide skin care suggestion according to skin condition, need to process facial image, quantification is glossy, pore these skin features thick, also lacking special image-recognizing method at present can accomplish these.
Summary of the invention
For the problem of these image recognitions above, this invention takes a kind of quantification skin quality detection method based on facial image identification, wherein have employed various characteristics of image recognition methods, the basis of recognition of face achieves quantification skin quality and detects.
The technical solution used in the present invention adopts following methods step:
1) by acquisition two human face photo clearly of taking pictures, one opens and opens flashlamp, another clear picture but do not open flashlamp, carries out recognition of face to human face photo, and the human face obtained after identifying is divided into eight face-image blocks;
The requirement of the present invention to facial image is: not wearing spectacles, and facial zone does not block by hair; Shooting at close range, face position is clear.Take two photos under these conditions, wherein one opens and opens flashlamp, and another Zhang Guangzhao well but do not open flashlamp.The basis that image is qualified adopts the methods such as recognition of face, Region dividing, region recognition successively, determines the skin quality situation of human face in picture.
2) to step 1) divide after each face-image block carry out the calculating of skin attribute;
3) according to the result of each block skin attribute, detect judgement and obtain skin quality type.
Described step 1) in human face photo, the face of people does not block by hair, and face position is clear.
Described step 1) in, recognition of face adopts third party's recognition of face instrument face++ to carry out recognition of face, obtains the following key point of people face, and marks off block:
1.1) key point comprises and is arranged in left key point on the left of face, is positioned at the right key point on the right side of face and is positioned at the key point in the middle of face:
Left key point comprises left volume point, left eyebrow right hand edge point, left eyebrow left hand edge point, left eye base point, left wing of nose point, left cheekbone point, middle left hand edge point, lower lip down contour point, bottom left marginal point, left corners of the mouth point and lower-left bar point; Middle key point comprises point in the middle part of left volume summit, right volume, right eye left hand edge and bottom left wing of nose point; Right key point comprises right volume point, right eyebrow right hand edge point, right eyebrow right hand edge point, right eye base point, right wing of nose point, right cheekbone point, middle right hand edge point, lower lip down contour point, bottom right marginal point, right corners of the mouth point and bottom right bar point;
1.2) respectively as two of rectangle block, each face-image block is determined to angle point by two key points: the face-image block on the left of face divides:
Face-image block L1 is marked off by left volume point and left eyebrow right hand edge point, face-image block L2 is marked off by left anchor point C and left wing of nose point, the horizontal ordinate of left anchor point C is identical with the horizontal ordinate of left eyebrow left hand edge point, and the ordinate of left anchor point C is identical with the ordinate of left eye base point;
Face-image block in the middle of face divides: mark off face-image block M1 by point and left eyebrow left hand edge point in the middle part of right volume, face-image block M2 is marked off by left eyebrow right hand edge point and right eye left hand edge point, mark off face-image block M3 by left eye right hand edge point, right eye left hand edge point and lower prenasale, mark off image block M4 by lower lip down contour point and left and right chin point;
Face-image block R1 is marked off by right volume point and right eyebrow right hand edge point, face-image block R2 is marked off by right anchor point B and right wing of nose point, the horizontal ordinate of right anchor point B is identical with the horizontal ordinate of right eyebrow right hand edge point, and the ordinate of right anchor point B is identical with the ordinate of right eye base point.
For described step 2) each face-image block carries out the calculating of skin attribute in the following ways:
2.2.1) glossy degree score value Oiliness is calculated for the human face photo septum reset image block M3 opening the human face photo septum reset image block L1 of flashlamp, L2, R1, R2, M1, M2, M4 and do not open flashlamp, and on average obtain the glossy degree score value in facial T district
with the glossy degree score value of cheek
2.2.2) for human face photo septum reset image block L1, L2, R1, R2, M1 and M3 of opening flashlamp, calculate skin smooth degree score value Smoothness, be then averaged and obtain the skin smooth degree score value in facial T district
with the skin smooth degree score value of cheek
2.2.3) for the human face photo septum reset image block L2, R2 and M3 that open flashlamp, calculate skin pore thick degree score value Pore, be then averaged the thick degree score value of skin pore obtaining facial T district
degree score value thick with the skin pore of cheek
2.2.4) for the human face photo septum reset image block M3 opening flashlamp, total blackhead order of severity score value Blackhead is calculated.
Described step 2.2.1) specifically calculate in the following ways: first adopt reflective recognition methods to remove the reflective of face-image block by original facial image block, by original facial image block and go reflective after the face-image block that obtains all transfer gray-scale map to, each pixel of original facial image block is traveled through, the pixel finding out original facial image block with go reflective after face-image block between pixel corresponding to position grey value difference be greater than 40 and the grey scale pixel value of this original facial image block all pixels of being greater than 140, obtain its total area, divided by the image block total area, obtain glossy area ratio y, then: as glossy area ratio y=0, then the glossy degree score value Oiliness of this face-image block is 0, as glossy area ratio y>0, then the glossy degree score value Oiliness=10+log of this face-image block
3.5s,
Calculate the glossy degree score value of mean value as this human face T district of the glossy degree score value Oiliness of face-image block L1, R1, M1, M2, M3 and M4 again
calculate the mean value of the glossy degree score value Oiliness of face-image block L2 and R2 as the glossy degree score value obtaining this person's cheek
the glossy degree that glossy degree score value Oiliness causes because of sebum secretion for embodying skin.
Described step 2.2.2) specifically calculate in the following ways: first use median filter to carry out filtering and noise reduction to original facial image block, re-use local auto-adaptive threshold method (Local adaptivethresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain; The connected domain quantity of original facial image block is designated as n, and the connected domain quantity of filtering and noise reduction process rear face image block is designated as dn, then: as dn=0, and smooth degree score value Smoothness=10; As dn>0, smooth degree score value Smoothness=140-(n – dn) 2/dn;
Calculate the skin smooth degree score value of mean value as this human face T district of the smooth degree score value Smoothness of face-image block L1, R1, M1 and M3 again
calculate the skin smooth degree score value of mean value as this person's cheek of the smooth degree score value Smoothness of face-image block L2 and R2
Described step 2.2.3) specifically calculate in the following ways: first use median filter to carry out filtering and noise reduction to original facial image block, re-use local auto-adaptive threshold method (Local adaptive thresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-componentlabeling) to calculate the quantity of wherein connected domain and respective area; Filter out area between 4 × 10
-5times original facial image block area A rea and 6 × 10
-4connected domain between times original facial image block area A rea is detected as pore part, the total area of all pore parts obtains pore area ratio m divided by original facial image block area A rea, calculates pore thick degree score value Pore=m × 1000 of this face-image block; Using the skin pore thick degree score value of the pore thick degree score value Pore of face-image block M3 as this human face T district
calculate the skin pore thick degree score value of mean value as this person's cheek of the pore thick degree score value Pore of face-image block L2 and R2
Described step 2.2.4) specifically calculate in the following ways: original facial image block is converted into gray-scale map, re-use local auto-adaptive threshold method (Local adaptive thresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain and respective area; Filter out area between 7 × 10
-5times original facial image block area A rea and 1 × 10
-3between times original facial image block area A rea and the connected domain that mean pixel gray-scale value is less than 180 is detected as blackhead part; The blackhead order of severity parameter of single blackhead part is the area of (the mean pixel gray-scale value of this blackhead part of 255-) × this blackhead part, be added by the blackhead order of severity parameter of each blackhead part and obtain total blackhead order of severity parameter h divided by original facial image block area A rea again, then calculate total blackhead order of severity score value Blackhead=log
1.4(2.5 × h+0.4).
Described step 3) specifically carry out calculating skin oiliness degree value GT and skin oiliness degree GC in the following ways and carry out judgement detecting:
3.1) the skin oiliness degree value GT in following formulae discovery face T district is adopted:
As GT<0.5, the skin detection in facial T district is dryness;
As 0.5<=GT<3.0, the skin detection in facial T district is neutral;
As GT>=3.0, the skin detection in facial T district is oiliness;
3.2) the skin oiliness degree GC of following formulae discovery two cheek is adopted:
As GC<0.5, the skin detection of cheek is dryness; As 0.5<=GC<3.0, the skin detection of cheek is neutral; As GC>=3.0, the skin detection of cheek is oiliness.
Described facial T district is made up of face-image block L1, R1, M1, M2, M3 and M4.
For common people, often comparatively cheek skin is slightly oily in T district, thus the present invention using T district skin as basic criterion: when T district skin is dryness, judge that overall face skin is as dryness; When T district skin is neutral, the situation according to cheek skin judges overall face skin as neutrality or neutrality partially dry; When T district skin is oiliness, the situation according to cheek skin judges that overall face skin is as oiliness or Combination.So can obtain neutral skin quality common in skin nursing, dry skin, oiliness skin quality, the classification of Combination skin quality.
The invention has the beneficial effects as follows:
The present invention proposes automatic ration from face picture innovatively and detects the serial of methods identifying skin quality oiliness degree, can meet the requirements to illumination, picture clearly human face photo resolve, gone out the skin skin quality of a people by the Face datection in image, thus next step can provide the suggestion giving optimal skin care method and skin care item.
The present invention can be widely used in beauty and skin care field, selects by people being helped the identification of facial image whenever and wherever possible, quickly and easily to make correct beauty and skin care with detection.
Accompanying drawing explanation
Fig. 1 is the division schematic diagram of face-image block of the present invention.
Fig. 2 is the human face photo that the embodiment of the present invention opens flashlamp.
Fig. 3 be embodiment of the present invention illumination well but do not open the human face photo of flashlamp.
Embodiment
Below in conjunction with drawings and the specific embodiments, the present invention is described in further detail.
1., based on a quantification skin quality detection method for facial image identification, it is characterized in that:
1) by acquisition two human face photo clearly of taking pictures, one opens and opens flashlamp, the good clear picture of another Zhang Guangzhao but do not open flashlamp, carries out recognition of face to human face photo, and the human face obtained after identifying is divided into eight face-image blocks; The good clear picture of above-mentioned illumination refers to that shadow-free and pore and hair are clear and legible on the face;
2) to step 1) divide after each face-image block carry out the calculating of skin attribute;
3) according to the result of each block skin attribute, detect judgement and obtain skin quality type.
Step 1) in human face photo, the face of people does not block by hair, and face position is clear, preferably at shooting at close range.
Step 1) in, recognition of face adopts third party's recognition of face instrument face++ to carry out recognition of face, obtains the following key point of people face, and marks off block:
1.1) key point comprises and is arranged in left key point on the left of face, is positioned at the right key point on the right side of face and is positioned at the key point in the middle of face:
Left key point comprises left volume point, left eyebrow right hand edge point, left eyebrow left hand edge point, left eye base point, left wing of nose point, left cheekbone point, middle left hand edge point, lower lip down contour point, bottom left marginal point, left corners of the mouth point and lower-left bar point;
Middle key point comprises point in the middle part of left volume summit, right volume, right eye left hand edge and bottom left wing of nose point;
Right key point comprises right volume point, right eyebrow right hand edge point, right eyebrow right hand edge point, right eye base point, right wing of nose point, right cheekbone point, middle right hand edge point, lower lip down contour point, bottom right marginal point, right corners of the mouth point and bottom right bar point;
1.2) respectively as two of rectangle block, each face-image block is determined to angle point by two key points: the face-image block on the left of face divides:
Face-image block L1 is marked off by left volume point and left eyebrow right hand edge point, face-image block L2 is marked off by left anchor point C and left wing of nose point, the horizontal ordinate of left anchor point C is identical with the horizontal ordinate of left eyebrow left hand edge point, and the ordinate of left anchor point C is identical with the ordinate of left eye base point; Face-image block in the middle of face divides: mark off face-image block M1 by point and left eyebrow left hand edge point in the middle part of right volume, face-image block M2 is marked off by left eyebrow right hand edge point and right eye left hand edge point, mark off face-image block M3 by left eye right hand edge point, right eye left hand edge point and lower prenasale, mark off image block M4 by lower lip down contour point and left and right chin point; Obtain by the above mode identical with on the left of face the right key point be positioned on the right side of face and mark off face-image block R1 and R2, face-image block R1 is marked off by right volume point and right eyebrow right hand edge point, face-image block R2 is marked off by right anchor point B and right wing of nose point, the horizontal ordinate of right anchor point B is identical with the horizontal ordinate of right eyebrow right hand edge point, and the ordinate of right anchor point B is identical with the ordinate of right eye base point.According to everyone face position feature, may there is overlap in above-mentioned block.
Each face-image block described is carried out in the following ways to the calculating of skin attribute:
2.2.1) glossy degree score value Oiliness is calculated for the human face photo septum reset image block M3 opening the human face photo septum reset image block L1 of flashlamp, L2, R1, R2, M1, M2, M4 and do not open flashlamp, and on average obtain the glossy degree score value in facial T district
with the glossy degree score value of cheek
Reflective recognition methods is first adopted to remove the reflective of face-image block by original facial image block, by original facial image block and go reflective after the face-image block that obtains all transfer gray-scale map to, each pixel of original facial image block is traveled through, the pixel finding out original facial image block with go reflective after face-image block between pixel corresponding to position grey value difference be greater than 40 and the grey scale pixel value of this original facial image block all pixels of being greater than 140, obtain its total area, divided by the image block total area, obtain glossy area ratio y, then:
As glossy area ratio y=0, then the glossy degree score value Oiliness of this face-image block is 0;
As glossy area ratio y>0, then the glossy degree score value Oiliness=10+log of this face-image block
3.5s;
Calculate the glossy degree score value of mean value as this human face T district of the glossy degree score value Oiliness of face-image block L1, R1, M1, M2, M3 and M4
calculate the mean value of the glossy degree score value Oiliness of face-image block L2 and R2 as the glossy degree score value obtaining this person's cheek
the glossy degree that glossy degree score value Oiliness causes because of sebum secretion for embodying skin.
Reflective recognition methods is the reflective recognition methods in paper Efficient and Robust Specular Highlight Removal (Q.Yang, 2014).The process employs widely used diffuse reflection+specular reflectance model in computer graphical; this model using specular reflection component as noise, so facility image denoising mode removes mirror-reflection Gao Guang: remove mirror-reflection chromatic component by edge-protected low-pass filter process original color image.
2.2.2) for human face photo septum reset image block L1, L2, R1, R2, M1 and M3 of opening flashlamp, calculate skin smooth degree score value Smoothness, be then averaged and obtain the skin smooth degree score value in facial T district
with the skin smooth degree score value of cheek
Median filter is first used to carry out filtering and noise reduction to original facial image block, re-use local auto-adaptive threshold method (Local adaptive thresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain; The connected domain quantity of original facial image block is designated as n, and the connected domain quantity of filtering and noise reduction process rear face image block is designated as dn, then:
As dn=0, smooth degree score value Smoothness=10;
As dn>0, smooth degree score value Smoothness=140-(n – dn)
2/ dn;
Calculate the skin smooth degree score value of mean value as this human face T district of the smooth degree score value Smoothness of face-image block L1, R1, M1 and M3
calculate the skin smooth degree score value of mean value as this person's cheek of the smooth degree score value Smoothness of face-image block L2 and R2
2.2.3) for the human face photo septum reset image block L2, R2 and M3 that open flashlamp, calculate skin pore thick degree score value Pore, be then averaged the thick degree score value of skin pore obtaining facial T district
degree score value thick with the skin pore of cheek
Median filter is first used to carry out filtering and noise reduction to original facial image block, re-use local auto-adaptive threshold method (Local adaptive thresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain and respective area; Filter out area between 4 × 10
-5times original facial image block area A rea and 6 × 10
-4connected domain between times original facial image block area A rea is detected as pore part, the total area of all pore parts obtains pore area ratio m divided by original facial image block area A rea, calculates pore thick degree score value Pore=m × 1000 of this face-image block; Using the skin pore thick degree score value of the pore thick degree score value Pore of face-image block M3 as this human face T district
calculate the skin pore thick degree score value of mean value as this person's cheek of the pore thick degree score value Pore of face-image block L2 and R2
2.2.4) for the human face photo septum reset image block M3 opening flashlamp, total blackhead order of severity score value Blackhead is calculated.
Original facial image block is converted into gray-scale map, re-use local auto-adaptive threshold method (Localadaptive thresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain and respective area; Filter out area between 7 × 10
-5times original facial image block area A rea and 1 × 10
-3between times original facial image block area A rea and the connected domain that mean pixel gray-scale value is less than 180 is detected as blackhead part; The blackhead order of severity parameter of single blackhead part is the area of (the mean pixel gray-scale value of this blackhead part of 255-) × this blackhead part, be added by the blackhead order of severity parameter of each blackhead part and obtain total blackhead order of severity parameter h divided by original facial image block area A rea again, then calculate total blackhead order of severity score value Blackhead=log
1.4(2.5 × h+0.4).
2.2.5) comprehensive above-mentioned result of calculation calculates skin oiliness degree value GT and skin oiliness degree GC and carries out judgement and detects.
Adopt the skin oiliness degree value GT in following formulae discovery face T district:
As GT<0.5, the skin detection in facial T district is dryness; As 0.5<=GT<3.0, the skin detection in facial T district is neutral; As GT>=3.0, the skin detection in facial T district is oiliness;
Adopt the skin oiliness degree GC of following formulae discovery two cheek again:
As GC<0.5, the skin detection of cheek is dryness; As 0.5<=GC<3.0, the skin detection of cheek is neutral; As GC>=3.0, the skin detection of cheek is oiliness.
Above-mentioned facial T district is made up of face-image block L1, R1, M1, M2, M3 and M4.
Embodiments of the invention are as follows:
A) input the facial image shown in Fig. 2 and Fig. 3, and its face is divided into 8 blocks;
B) calculate the glossy degree of skin quality, the score value obtaining the Oiliness of L1, R1, M1, M2, M3, M4 is followed successively by 7.35,7.13,7.60,6.29,8.29,3.21, on average obtains this glossy degree score value in human face T district
be 6.65.The score value of L2, R2 is respectively 6.51,5.02, on average obtains the glossy degree score value of cheek
be 5.77.
C) calculate skin quality smooth degree, the Smoothness score value of L1, R1, M1, M3 is respectively 8.35,5.24,0.12,8.28, obtains this human face T district skin quality smooth degree score value
be 5.49, L2, the score value of R2 is respectively 8.67,7.68, obtain cheek smooth degree score value
be 8.17.
D) calculate the thick degree of pore, the Pore score value obtaining M3 is 5.84, is the thick degree score value of this human face T district pore
the score value Pore score value of L2, R2 is respectively 7.65,7.42, namely obtains the thick degree score value of this person's cheek pore
7.54.
E) calculate the blackhead order of severity, obtain blackhead order of severity score value Blackhead=1.02.
F) gather above-mentioned result of calculation, obtain GT=5.12, T district skin is oiliness.
GC=4.81, cheek skin is oiliness.Judge that overall face skin is as oiliness.
G) after obtaining his skin quality information, if current sensing time is summer, temperature, humidity are higher, illumination is strong, so just he can be advised: 1. sooner or later use clean of amino acid pattern mildy wash, note clean, moisturizing and the control oil of face, especially forehead and nose, prevents from growing dark sore and small pox.2. suitably using man's toner and man's facial mask, can pore refining be played, removing the effect of blackhead.If current sensing time is autumn, temperature humidity is lower, and illumination is not strong, then only need to carry out clean sooner or later.
Above-mentioned embodiment is used for explaining and the present invention is described, instead of limits the invention, and in the protection domain of spirit of the present invention and claim, any amendment make the present invention and change, all fall into protection scope of the present invention.
Claims (10)
1., based on a quantification skin quality detection method for facial image identification, it is characterized in that:
1) by acquisition two human face photo clearly of taking pictures, one opens and opens flashlamp, another clear picture but do not open flashlamp, carries out recognition of face to human face photo, and the human face obtained after identifying is divided into eight face-image blocks;
2) to step 1) divide after each face-image block carry out the calculating of skin attribute;
3) according to the result of each block skin attribute, detect judgement and obtain skin quality type.
2. a kind of quantification skin quality detection method based on facial image identification according to claim 1, is characterized in that: described step 1) in human face photo, the face of people does not block by hair, and face position is clear.
3. a kind of quantification skin quality detection method based on facial image identification according to claim 1, it is characterized in that: described step 1) in, recognition of face adopts third party's recognition of face instrument face++ to carry out recognition of face, obtains the following key point of people face, and marks off block:
1.1) key point comprises and is arranged in left key point on the left of face, is arranged in the right key point on the right side of face and is positioned at the key point in the middle of face: left key point comprises left volume point, left eyebrow right hand edge point, left eyebrow left hand edge point, left eye base point, left wing of nose point, left cheekbone point, left hand edge point, lower lip down contour point, bottom left marginal point, left corners of the mouth point and lower-left bar point; Middle key point comprises point in the middle part of left volume summit, right volume, right eye left hand edge and bottom left wing of nose point; Right key point comprises right volume point, right eyebrow right hand edge point, right eyebrow right hand edge point, right eye base point, right wing of nose point, right cheekbone point, middle right hand edge point, lower lip down contour point, bottom right marginal point, right corners of the mouth point and bottom right bar point;
1.2) respectively as two of rectangle block, each face-image block is determined to angle point by two key points: the face-image block on the left of face divides:
Face-image block L1 is marked off by left volume point and left eyebrow right hand edge point, face-image block L2 is marked off by left anchor point C and left wing of nose point, the horizontal ordinate of left anchor point C is identical with the horizontal ordinate of left eyebrow left hand edge point, and the ordinate of left anchor point C is identical with the ordinate of left eye base point;
Face-image block in the middle of face divides: mark off face-image block M1 by point and left eyebrow left hand edge point in the middle part of right volume, face-image block M2 is marked off by left eyebrow right hand edge point and right eye left hand edge point, mark off face-image block M3 by left eye right hand edge point, right eye left hand edge point and lower prenasale, mark off image block M4 by lower lip down contour point and left and right chin point;
Face-image block R1 is marked off by right volume point and right eyebrow right hand edge point, face-image block R2 is marked off by right anchor point B and right wing of nose point, the horizontal ordinate of right anchor point B is identical with the horizontal ordinate of right eyebrow right hand edge point, and the ordinate of right anchor point B is identical with the ordinate of right eye base point.
4. a kind of quantification skin quality detection method based on facial image identification according to claim 1, is characterized in that: for described step 2) each face-image block carries out the calculating of skin attribute in the following ways:
2.2.1) glossy degree score value Oiliness is calculated for the human face photo septum reset image block M3 opening the human face photo septum reset image block L1 of flashlamp, L2, R1, R2, M1, M2, M4 and do not open flashlamp, and on average obtain the glossy degree score value in facial T district
with the glossy degree score value of cheek
2.2.2) for human face photo septum reset image block L1, L2, R1, R2, M1 and M3 of opening flashlamp, calculate skin smooth degree score value Smoothness, be then averaged and obtain the skin smooth degree score value in facial T district
with the skin smooth degree score value of cheek
2.2.3) for the human face photo septum reset image block L2, R2 and M3 that open flashlamp, calculate skin pore thick degree score value Pore, be then averaged the thick degree score value of skin pore obtaining facial T district
degree score value thick with the skin pore of cheek
2.2.4) for the human face photo septum reset image block M3 opening flashlamp, total blackhead order of severity score value Blackhead is calculated.
5. a kind of quantification skin quality detection method based on facial image identification according to claim 4, it is characterized in that: described step 2.2.1) specifically calculate in the following ways: first adopt reflective recognition methods to remove the reflective of face-image block by original facial image block, by original facial image block and go reflective after the face-image block that obtains all transfer gray-scale map to, each pixel of original facial image block is traveled through, the pixel finding out original facial image block with go reflective after face-image block between pixel corresponding to position grey value difference be greater than 40 and the grey scale pixel value of this original facial image block all pixels of being greater than 140, obtain its total area, divided by the image block total area, obtain glossy area ratio y, then: as glossy area ratio y=0, then the glossy degree score value Oiliness of this face-image block is 0, as glossy area ratio y>0, then the glossy degree score value Oiliness=10+log of this face-image block
3.5s,
Calculate the glossy degree score value of mean value as this human face T district of the glossy degree score value Oiliness of face-image block L1, R1, M1, M2, M3 and M4 again
calculate the mean value of the glossy degree score value Oiliness of face-image block L2 and R2 as the glossy degree score value obtaining this person's cheek
the glossy degree that glossy degree score value Oiliness causes because of sebum secretion for embodying skin.
6. a kind of quantification skin quality detection method based on facial image identification according to claim 4, it is characterized in that: described step 2.2.2) specifically calculate in the following ways: first use median filter to carry out filtering and noise reduction to original facial image block, re-use local auto-adaptive threshold method (Local adaptivethresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain; The connected domain quantity of original facial image block is designated as n, and the connected domain quantity of filtering and noise reduction process rear face image block is designated as dn, then:
As dn=0, smooth degree score value Smoothness=10; As dn>0, smooth degree score value Smoothness=140-(n – dn)
2/ dn;
Calculate the skin smooth degree score value of mean value as this human face T district of the smooth degree score value Smoothness of face-image block L1, R1, M1 and M3 again
calculate the skin smooth degree score value of mean value as this person's cheek of the smooth degree score value Smoothness of face-image block L2 and R2
7. a kind of quantification skin quality detection method based on facial image identification according to claim 4, it is characterized in that: described step 2.2.3) specifically calculate in the following ways: first use median filter to carry out filtering and noise reduction to original facial image block, re-use local auto-adaptive threshold method (Local adaptivethresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain and respective area; Filter out area between 4 × 10
-5times original facial image block area A rea and 6 × 10
-4connected domain between times original facial image block area A rea is detected as pore part, the total area of all pore parts obtains pore area ratio m divided by original facial image block area A rea, calculates pore thick degree score value Pore=m × 1000 of this face-image block; Again using the skin pore thick degree score value of the pore thick degree score value Pore of face-image block M3 as this human face T district
calculate the skin pore thick degree score value of mean value as this person's cheek of the pore thick degree score value Pore of face-image block L2 and R2
8. a kind of quantification skin quality detection method based on facial image identification according to claim 4, it is characterized in that: described step 2.2.4) specifically calculate in the following ways: original facial image block is converted into gray-scale map, re-use local auto-adaptive threshold method (Local adaptive thresholding) by binaryzation in face-image block, then use connected domain analysis method (Connected-component labeling) to calculate the quantity of wherein connected domain and respective area; Filter out area between 7 × 10
-5times original facial image block area A rea and 1 × 10
-3between times original facial image block area A rea and the connected domain that mean pixel gray-scale value is less than 180 is detected as blackhead part; The blackhead order of severity parameter of single blackhead part is the area of (the mean pixel gray-scale value of this blackhead part of 255-) × this blackhead part, be added by the blackhead order of severity parameter of each blackhead part and obtain total blackhead order of severity parameter h divided by original facial image block area A rea again, then calculate total blackhead order of severity score value Blackhead=log
1.4(2.5 × h+0.4).
9. a kind of quantification skin quality detection method based on facial image identification according to claim 1, is characterized in that: described step 3) specifically carry out calculating skin oiliness degree value GT and skin oiliness degree GC in the following ways and carry out judgement detecting:
3.1) the skin oiliness degree value GT in following formulae discovery face T district is adopted:
As GT<0.5, the skin detection in facial T district is dryness;
As 0.5<=GT<3.0, the skin detection in facial T district is neutral;
As GT>=3.0, the skin detection in facial T district is oiliness;
3.2) the skin oiliness degree GC of following formulae discovery two cheek is adopted:
As GC<0.5, the skin detection of cheek is dryness; As 0.5<=GC<3.0, the skin detection of cheek is neutral; As GC>=3.0, the skin detection of cheek is oiliness.
10., according to claim 4 ~ 7 or 9 arbitrary described a kind of quantification skin quality detection methods based on facial image identification, it is characterized in that: described facial T district is made up of face-image block L1, R1, M1, M2, M3 and M4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510130557.1A CN104732214B (en) | 2015-03-24 | 2015-03-24 | A kind of quantification skin quality detection method based on facial image identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510130557.1A CN104732214B (en) | 2015-03-24 | 2015-03-24 | A kind of quantification skin quality detection method based on facial image identification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104732214A true CN104732214A (en) | 2015-06-24 |
| CN104732214B CN104732214B (en) | 2018-10-12 |
Family
ID=53456088
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510130557.1A Active CN104732214B (en) | 2015-03-24 | 2015-03-24 | A kind of quantification skin quality detection method based on facial image identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104732214B (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105913389A (en) * | 2016-04-07 | 2016-08-31 | 广东欧珀移动通信有限公司 | Image processing method and device for skin abnormity |
| CN106295615A (en) * | 2016-08-23 | 2017-01-04 | 姜汇洋 | Skin detection based on mobile interchange terminal and auxiliary maintenance system |
| CN106469302A (en) * | 2016-09-07 | 2017-03-01 | 成都知识视觉科技有限公司 | A kind of face skin quality detection method based on artificial neural network |
| CN107157447A (en) * | 2017-05-15 | 2017-09-15 | 精诚工坊电子集成技术（北京）有限公司 | The detection method of skin surface roughness based on image RGB color |
| CN107437073A (en) * | 2017-07-19 | 2017-12-05 | 竹间智能科技（上海）有限公司 | Face skin quality analysis method and system based on deep learning with generation confrontation networking |
| CN107645636A (en) * | 2017-08-31 | 2018-01-30 | 维沃移动通信有限公司 | A kind of method taken pictures and mobile terminal |
| CN107692965A (en) * | 2016-08-09 | 2018-02-16 | 上海图檬信息科技有限公司 | The method of guiding and positioning skin detection equipment based on mobile phone camera to specified location |
| CN108323203A (en) * | 2017-07-17 | 2018-07-24 | 深圳和而泰智能控制股份有限公司 | A kind of method, apparatus and intelligent terminal quantitatively detecting face skin quality parameter |
| CN108369644A (en) * | 2017-07-17 | 2018-08-03 | 深圳和而泰智能控制股份有限公司 | A kind of method and intelligent terminal quantitatively detecting face wrinkles on one's forehead |
| CN108363965A (en) * | 2018-01-29 | 2018-08-03 | 杭州美界科技有限公司 | A kind of distributed face state appraisal procedure |
| CN108354590A (en) * | 2018-01-29 | 2018-08-03 | 杭州美界科技有限公司 | A kind of face state appraisal procedure based on burst mode |
| CN108389185A (en) * | 2018-01-29 | 2018-08-10 | 杭州美界科技有限公司 | A kind of face state appraisal procedure |
| CN108399364A (en) * | 2018-01-29 | 2018-08-14 | 杭州美界科技有限公司 | A kind of face state appraisal procedure of major-minor camera setting |
| CN108553083A (en) * | 2018-01-29 | 2018-09-21 | 杭州美界科技有限公司 | A kind of face state appraisal procedure under voice instruction |
| CN108618752A (en) * | 2017-03-23 | 2018-10-09 | 丽宝大数据股份有限公司 | Skin product adaptation method and its electronic device |
| CN109300105A (en) * | 2017-07-25 | 2019-02-01 | 上海中科顶信医学影像科技有限公司 | Pore detection method, system, equipment and storage medium |
| CN109522839A (en) * | 2018-11-15 | 2019-03-26 | 北京达佳互联信息技术有限公司 | A kind of face skin area determines method, apparatus, terminal device and storage medium |
| WO2019072249A1 (en) * | 2017-10-12 | 2019-04-18 | 精诚工坊电子集成技术（北京）有限公司 | Method and device for analyzing water content of skin by means of skin image |
| WO2020015629A1 (en) * | 2018-07-20 | 2020-01-23 | 华为技术有限公司 | Image acquiring method and device, and terminal |
| CN111755119A (en) * | 2020-06-22 | 2020-10-09 | 南方医科大学 | Method, system, device and medium for analyzing facial skin data based on region division |
| WO2020207423A1 (en) * | 2019-04-12 | 2020-10-15 | 虹软科技股份有限公司 | Skin type detection method, skin type grade classification method and skin type detection apparatus |
| CN112037162A (en) * | 2019-05-17 | 2020-12-04 | 华为技术有限公司 | Facial acne detection method and equipment |
| WO2022037105A1 (en) * | 2020-08-17 | 2022-02-24 | 苏州瀚华智造智能技术有限公司 | Automatic braces cutting line generation method, and application |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1297782A1 (en) * | 2001-10-01 | 2003-04-02 | L'oreal | Beauty analysis of external body conditions |
| CN101065061A (en) * | 2004-10-29 | 2007-10-31 | 强生消费者公司 | Apparatus for and method of taking and viewing images of the skin |
| CN104299011A (en) * | 2014-10-13 | 2015-01-21 | 吴亮 | Skin type and skin problem identification and detection method based on facial image identification |
-
2015
- 2015-03-24 CN CN201510130557.1A patent/CN104732214B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1297782A1 (en) * | 2001-10-01 | 2003-04-02 | L'oreal | Beauty analysis of external body conditions |
| CN101065061A (en) * | 2004-10-29 | 2007-10-31 | 强生消费者公司 | Apparatus for and method of taking and viewing images of the skin |
| CN104299011A (en) * | 2014-10-13 | 2015-01-21 | 吴亮 | Skin type and skin problem identification and detection method based on facial image identification |
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105913389B (en) * | 2016-04-07 | 2018-05-04 | 广东欧珀移动通信有限公司 | Image processing method and device during skin abnormality |
| CN105913389A (en) * | 2016-04-07 | 2016-08-31 | 广东欧珀移动通信有限公司 | Image processing method and device for skin abnormity |
| CN107692965A (en) * | 2016-08-09 | 2018-02-16 | 上海图檬信息科技有限公司 | The method of guiding and positioning skin detection equipment based on mobile phone camera to specified location |
| CN106295615A (en) * | 2016-08-23 | 2017-01-04 | 姜汇洋 | Skin detection based on mobile interchange terminal and auxiliary maintenance system |
| CN106469302A (en) * | 2016-09-07 | 2017-03-01 | 成都知识视觉科技有限公司 | A kind of face skin quality detection method based on artificial neural network |
| CN106469302B (en) * | 2016-09-07 | 2019-05-28 | 成都知识视觉科技有限公司 | A kind of face skin quality detection method based on artificial neural network |
| CN108618752A (en) * | 2017-03-23 | 2018-10-09 | 丽宝大数据股份有限公司 | Skin product adaptation method and its electronic device |
| CN107157447A (en) * | 2017-05-15 | 2017-09-15 | 精诚工坊电子集成技术（北京）有限公司 | The detection method of skin surface roughness based on image RGB color |
| CN107157447B (en) * | 2017-05-15 | 2020-03-20 | 北京工商大学 | Skin surface roughness detection method based on image RGB color space |
| WO2019014813A1 (en) * | 2017-07-17 | 2019-01-24 | 深圳和而泰智能控制股份有限公司 | Method and apparatus for quantitatively detecting skin type parameter of human face, and intelligent terminal |
| CN108323203A (en) * | 2017-07-17 | 2018-07-24 | 深圳和而泰智能控制股份有限公司 | A kind of method, apparatus and intelligent terminal quantitatively detecting face skin quality parameter |
| CN108369644B (en) * | 2017-07-17 | 2022-05-17 | 深圳和而泰智能控制股份有限公司 | Method for quantitatively detecting human face raised line, intelligent terminal and storage medium |
| WO2019014814A1 (en) * | 2017-07-17 | 2019-01-24 | 深圳和而泰智能控制股份有限公司 | Method for quantitatively detecting forehead wrinkles on human face, and intelligent terminal |
| CN108369644A (en) * | 2017-07-17 | 2018-08-03 | 深圳和而泰智能控制股份有限公司 | A kind of method and intelligent terminal quantitatively detecting face wrinkles on one's forehead |
| CN107437073A (en) * | 2017-07-19 | 2017-12-05 | 竹间智能科技（上海）有限公司 | Face skin quality analysis method and system based on deep learning with generation confrontation networking |
| CN109300105A (en) * | 2017-07-25 | 2019-02-01 | 上海中科顶信医学影像科技有限公司 | Pore detection method, system, equipment and storage medium |
| CN107645636A (en) * | 2017-08-31 | 2018-01-30 | 维沃移动通信有限公司 | A kind of method taken pictures and mobile terminal |
| CN107645636B (en) * | 2017-08-31 | 2019-11-29 | 维沃移动通信有限公司 | A kind of method taken pictures and mobile terminal |
| CN109671046A (en) * | 2017-10-12 | 2019-04-23 | 精诚工坊电子集成技术（北京）有限公司 | Utilize the method and device of skin image analyzing skin moisture |
| US11490854B2 (en) * | 2017-10-12 | 2022-11-08 | Jingcheng Workshop Electronic Integration Technology (Beijing) Co., Ltd. | Method and device for analyzing water content of skin by means of skin image |
| WO2019072249A1 (en) * | 2017-10-12 | 2019-04-18 | 精诚工坊电子集成技术（北京）有限公司 | Method and device for analyzing water content of skin by means of skin image |
| CN108389185A (en) * | 2018-01-29 | 2018-08-10 | 杭州美界科技有限公司 | A kind of face state appraisal procedure |
| CN108354590A (en) * | 2018-01-29 | 2018-08-03 | 杭州美界科技有限公司 | A kind of face state appraisal procedure based on burst mode |
| CN108553083A (en) * | 2018-01-29 | 2018-09-21 | 杭州美界科技有限公司 | A kind of face state appraisal procedure under voice instruction |
| CN108363965A (en) * | 2018-01-29 | 2018-08-03 | 杭州美界科技有限公司 | A kind of distributed face state appraisal procedure |
| CN108399364A (en) * | 2018-01-29 | 2018-08-14 | 杭州美界科技有限公司 | A kind of face state appraisal procedure of major-minor camera setting |
| WO2020015629A1 (en) * | 2018-07-20 | 2020-01-23 | 华为技术有限公司 | Image acquiring method and device, and terminal |
| US11810277B2 (en) | 2018-07-20 | 2023-11-07 | Huawei Technologies Co., Ltd. | Image acquisition method, apparatus, and terminal |
| CN109522839A (en) * | 2018-11-15 | 2019-03-26 | 北京达佳互联信息技术有限公司 | A kind of face skin area determines method, apparatus, terminal device and storage medium |
| WO2020207423A1 (en) * | 2019-04-12 | 2020-10-15 | 虹软科技股份有限公司 | Skin type detection method, skin type grade classification method and skin type detection apparatus |
| CN112037162A (en) * | 2019-05-17 | 2020-12-04 | 华为技术有限公司 | Facial acne detection method and equipment |
| CN112037162B (en) * | 2019-05-17 | 2022-08-02 | 荣耀终端有限公司 | Facial acne detection method and equipment |
| CN111755119A (en) * | 2020-06-22 | 2020-10-09 | 南方医科大学 | Method, system, device and medium for analyzing facial skin data based on region division |
| CN111755119B (en) * | 2020-06-22 | 2024-04-05 | 南方医科大学 | Facial skin data analysis method, system, equipment and medium based on region division |
| WO2022037105A1 (en) * | 2020-08-17 | 2022-02-24 | 苏州瀚华智造智能技术有限公司 | Automatic braces cutting line generation method, and application |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104732214B (en) | 2018-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104732214A (en) | Quantification skin detecting method based on face image recognition | |
| CN104299011A (en) | Skin type and skin problem identification and detection method based on facial image identification | |
| CN104834898B (en) | A kind of quality classification method of personage's photographs | |
| KR100889014B1 (en) | Extraction method of tongue region using graph-based approach | |
| JP4307496B2 (en) | Facial part detection device and program | |
| JP4845698B2 (en) | Eye detection device, eye detection method, and program | |
| CA2751549C (en) | Method and apparatus for simulation of facial skin aging and de-aging | |
| JP5262243B2 (en) | Eye opening / closing discrimination device and program | |
| US20120155726A1 (en) | method and system of determining a grade of nuclear cataract | |
| CN111524080A (en) | Face skin feature identification method, terminal and computer equipment | |
| CN101339603A (en) | Method for selecting qualified iris image from video frequency stream | |
| CN107077596A (en) | System for producing the face-image met for selected identification document | |
| CN101447023B (en) | Method and system for detecting human head | |
| JP6822482B2 (en) | Line-of-sight estimation device, line-of-sight estimation method, and program recording medium | |
| CN101615241B (en) | Method for screening certificate photos | |
| JP2007272435A (en) | Face feature extraction device and face feature extraction method | |
| CN113012093B (en) | Training method and training system for glaucoma image feature extraction | |
| JP2022039984A (en) | System and method for finding area of eye from facial image | |
| JP3868089B2 (en) | Face determination method and face determination support device | |
| JP2019082743A (en) | Information processing apparatus and information processing method | |
| Tamilarasi et al. | Template matching algorithm for exudates detection from retinal fundus images | |
| Akhade et al. | Automatic optic disc detection in digital fundus images using image processing techniques | |
| Frame et al. | Texture analysis of retinal neovascularisation | |
| KR102282334B1 (en) | Method for optic disc classification | |
| Hosseini et al. | Facial expression analysis for estimating patient's emotional states in RPMS |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20181123 Address after: 311121 Room 303, 8 Building 1818-2 Wenyi West Road, Yuhang Street, Yuhang District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou to interest Internet Technology Co., Ltd. Address before: 322000 Wucun Group 3, Yiting Town, Yiwu City, Jinhua City, Zhejiang Province Patentee before: Wu Liang |
|
| TR01 | Transfer of patent right |