CN104424483A - Face image illumination preprocessing method, face image illumination preprocessing device and terminal - Google Patents
Face image illumination preprocessing method, face image illumination preprocessing device and terminal Download PDFInfo
- Publication number
- CN104424483A CN104424483A CN201310366699.9A CN201310366699A CN104424483A CN 104424483 A CN104424483 A CN 104424483A CN 201310366699 A CN201310366699 A CN 201310366699A CN 104424483 A CN104424483 A CN 104424483A
- Authority
- CN
- China
- Prior art keywords
- facial image
- illumination
- pretreatment
- several different
- parameter
- 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
- 238000005286 illumination Methods 0.000 title claims abstract description 227
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000007781 pre-processing Methods 0.000 title abstract description 10
- 230000001815 facial effect Effects 0.000 claims description 189
- 238000012545 processing Methods 0.000 claims description 19
- 238000005516 engineering process Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000009183 running Effects 0.000 description 4
- 239000000203 mixture Substances 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 241001497337 Euscorpius gamma Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/60—Extraction of image or video features relating to illumination properties, e.g. using a reflectance or lighting model
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a face image illumination preprocessing method comprising the steps of acquiring a face image and obtaining the illumination value of the environment where a terminal is disposed, looking up in a pre-stored face database a face image closest to the illumination value and obtaining preprocessing parameters corresponding to the face image closest to the illumination value, and carrying out illumination preprocessing on the acquired face image according to the preprocessing parameters. The invention further discloses a face image illumination preprocessing device and a terminal. By adopting the technical scheme of the invention, high-speed and high-recognition-rate face image illumination preprocessing can be carried out in an environment with changing illumination.
Description
Technical field
The present invention relates to the face image processing technology in recognition of face and light level technology, particularly relate to a kind of light irradiation preprocess method of facial image, device and terminal.
Background technology
The impact of illumination variation on recognition of face performance is one of terminal key issue being in mobile environment.In order to solve the impact of illumination variation on recognition of face performance, existing illumination pretreatment algorithm mainly does illumination adaptive based on facial image conversion and three-dimensional (3D, 3 Dimension) modeling and compensates.Facial image mapping algorithm in a mobile environment discrimination is poor, although 3D modeling algorithm has higher discrimination, needs to expend huge computing cost and time complexity.
When terminal is in mobile environment, the randomness of illumination variation is large, computing power is limited, if when existing illumination pretreatment algorithm operates in the terminal that is under mobile environment, processing speed and discrimination all can not get a desired effect.
Along with the impact being widely current and constantly living every day on people of intelligent terminal, the face recognition technology on intelligent terminal shows considerable marketable value and potentiality.But the restriction of processing speed and discrimination constrains the direct application of conventional face's recognition technology on intelligent terminal.Moreover, due to the ambulant feature of intelligent terminal, the recognition of face under mobile environment brings illumination variation, which increases the difficulty of mobile environment human face identification application.
Summary of the invention
In view of this, fundamental purpose of the present invention is to provide a kind of light irradiation preprocess method of facial image, device and terminal, in the environment of illumination variation, can carry out at a high speed and the illumination pretreatment of high discrimination facial image.
For achieving the above object, technical scheme of the present invention is achieved in that
A light irradiation preprocess method for facial image, described method comprises:
When gathering facial image, and obtain the illumination value of environment residing for terminal;
In the face database prestored, search facial image immediate with described illumination value, and obtain pretreatment parameter corresponding to the immediate facial image of described illumination value;
According to described pretreatment parameter, illumination pretreatment is carried out to the facial image collected.
Described method also comprises:
Gather facial image during multiple intensity of illumination respectively, and obtain the facial image of several different illumination intensity;
The illumination value of the facial image of several different illumination intensity described is obtained by described light sensor, and as reference illumination value;
The pretreatment parameter of the facial image of several different illumination intensity described is calculated respectively according to illumination pretreatment algorithm;
According to the pretreatment parameter of the facial image of several different illumination intensity described, the illumination value of the facial image of several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
When described illumination pretreatment algorithm is gamma intensity rectification GIC algorithm, described pretreatment parameter is gamma parameter.
Describedly in the face database prestored, search facial image immediate with described illumination value, and obtain pretreatment parameter corresponding to the immediate facial image of described illumination value, comprising:
Set up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
According to described discrete equation, calculate immediate with reference to illumination value with the illumination value of the facial image collected, and obtain the described pretreatment parameter with reference to illumination value.
Described method also comprises:
Recognition of face process is carried out to the facial image after illumination pretreatment.
The illumination pretreatment device of kind of facial image, described device comprises: collecting unit, acquiring unit, search unit and processing unit; Wherein,
Described collecting unit, during for gathering facial image, triggers described acquiring unit;
Described acquiring unit, for receive described collecting unit triggering after, obtain the illumination value of environment residing for terminal;
Describedly searching unit, for searching facial image immediate with described illumination value in the face database prestored, and obtaining pretreatment parameter corresponding to the immediate facial image of described illumination value;
Described processing unit, for carrying out illumination pretreatment according to described pretreatment parameter to the facial image collected.
Described device also comprises pretreatment unit and human face data library unit;
Described collecting unit, also for gathering facial image during multiple intensity of illumination respectively, and obtains the facial image of several different illumination intensity;
Described acquiring unit, also for obtaining the illumination value of the facial image of several different illumination intensity described, and as reference illumination value;
Described pretreatment unit, for calculating the pretreatment parameter of the facial image of several different illumination intensity described respectively according to illumination pretreatment algorithm;
Described human face data library unit, for the pretreatment parameter of the illumination value of the facial image of the facial image according to several different illumination intensity described, several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
When described illumination pretreatment algorithm is GIC algorithm, described pretreatment parameter is gamma parameter.
Described unit of searching comprises: set up subelement and computation subunit; Wherein,
Describedly set up subelement, for setting up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
Described computation subunit, for according to described discrete equation, calculates immediate with reference to illumination value with the illumination value of the facial image collected, and obtains the described pretreatment parameter with reference to illumination value.
Described device also comprises recognition unit, for carrying out recognition of face process to the facial image after illumination pretreatment.
A kind of terminal, described terminal comprises the illumination pretreatment device of above-mentioned any facial image.
The light irradiation preprocess method of facial image, device and terminal that the embodiment of the present invention is recorded, when gathering facial image, and obtain the illumination value of environment residing for terminal; In the face database prestored, search facial image immediate with described illumination value, and obtain pretreatment parameter corresponding to the immediate facial image of described illumination value; According to described pretreatment parameter, illumination pretreatment is carried out to the facial image collected.So, in the environment of illumination variation, facial image can be carried out at a high speed and the illumination pretreatment of high discrimination, thus while the high discrimination of guarantee, improve processing speed, reduce resource overhead.
Accompanying drawing explanation
Fig. 1 is the realization flow schematic diagram of the light irradiation preprocess method of embodiment of the present invention facial image;
Fig. 2 is the structure composition schematic diagram of the illumination pretreatment device of embodiment of the present invention facial image;
Fig. 3 is the structure composition schematic diagram searching unit in the illumination pretreatment device of embodiment of the present invention facial image;
Fig. 4 is the structure composition schematic diagram of embodiment of the present invention terminal;
Fig. 5 is the facial image schematic diagram in the embodiment of the present invention under different illumination conditions;
Fig. 6 is the facial image schematic diagram in the embodiment of the present invention under a sample different illumination conditions.
Embodiment
In order to more at large understand feature of the present invention and technology contents, below in conjunction with accompanying drawing, realization of the present invention is described in detail, the use of appended accompanying drawing explanation only for reference, is not used for limiting the present invention.
The basic thought of the embodiment of the present invention is: the illumination pretreatment of to carry out facial image by being embedded in data that light sensor on mobile phone faceplate gathers intensity of illumination as Study first, to reach in processing speed and the limited terminal being under mobile environment of discrimination, facial image is carried out at a high speed and the illumination pretreatment of high discrimination.
The embodiment of the present invention describes a kind of light irradiation preprocess method of facial image, as shown in Figure 1, said method comprising the steps of:
Step 101: when gathering facial image, and obtain the illumination value of environment residing for terminal.
Preferably, described method also comprises:
Gather facial image during multiple intensity of illumination respectively, and obtain the facial image of several different illumination intensity;
The illumination value of the facial image of several different illumination intensity described is obtained by described light sensor, and as reference illumination value;
The pretreatment parameter of the facial image of several different illumination intensity described is calculated respectively according to illumination pretreatment algorithm;
According to the pretreatment parameter of the facial image of several different illumination intensity described, the illumination value of the facial image of several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
Preferably, when described illumination pretreatment algorithm is gamma intensity rectification (GIC, Gamma Intensity Correction) algorithm, described pretreatment parameter is gamma parameter.
In order to clearly the embodiment of the present invention can be understood, now GIC algorithm is described in detail.
GIC algorithm, by the overall brightness of amendment facial image, makes this facial image and predefined standard faces image match.The facial image of standard normally lower of standard light conditions collects or revises the facial image obtained.
Such as, I is the facial image of input, and I ' is the facial image of I after GIC algorithm process, then I ' obtains by formula (1).
I'
xy=G(I
xy;γ
*) (1)
Wherein, I
xyrepresent that the facial image coordinate of input is the gray-scale value at (x, y) place; I'
xyrepresenting through G() conversion and the facial image coordinate that obtains be the gray-scale value at (x, y) place; G (I
xy; γ
*) represent I
xycarry out gamma transformation.
Gamma parameter γ in formula (1)
*formula (2) can be passed through calculate.Formula (2) for get optimum gamma parameter γ in multiple gamma parameter γ
*, the object of formula (2) finds one to have the gamma parameter γ of minimal difference with standard faces image
*.
Wherein,
Represent
Get γ value during minimum value; G (I
xy; Expression formula γ) is formula (3).
Wherein, c is gray scale warp parameter.
After said process, the facial image of several different illumination intensity can obtain corresponding pretreatment parameter, i.e. gamma parameter γ
*.
To mainly following two steps that time complexity contribution is maximum in above-mentioned GIC algorithm: 1) obtain gamma parameter γ
*optimization procedure; 2) with gamma parameter γ
*generate the process of last facial image.Further, except above-mentioned two processes, two parameters versus time complexities are also had to have important contribution, i.e. the scope of gamma parameter γ and search gamma parameter γ
*step size mu.Assuming that gamma parameter γ
*scope be formula (4),
γ
*∈[R
0,R
1],γ
*>0 (4)
Wherein, γ
*∈ [R
0, R
1] represent Suo Jiama parameter γ
*span for being greater than/equaling R
0and be less than etc./in R
1.
Suppose to obtain gamma parameter γ
*the time complexity of optimization procedure be T respectively
1, with gamma parameter γ
*the time complexity generating the process of last facial image is T
2.Then can obtain total time complexity T by following formula (5) to formula (9), at formula (5) in formula (9), O() be time complexity function, and, comprise n in face database and open facial image, often open the in the same size of facial image, H is the height of facial image, and W is the width of facial image.Below the computation process of total time complexity T is described in detail.
In conjunction with formula (4), the time complexity of formula (2) can be expressed as formula (5)
Wherein, the expression formula of Ψ is as formula (6).
Ψ=[G(I
xy;γ)-I
0(x,y)]
2(6)
T
1reduced form can pass through formula (7) and obtain.
Formula (1) utilizes the Suo Jiama parameter γ calculated
*produce final facial image, the account form of its time complexity as shown in Equation (8).
T
2=O(H·W) (8)
In sum, the account form of the All Time complexity of GIC algorithm is as formula (9).
Step 102: search facial image immediate with described illumination value in the face database prestored, and obtain pretreatment parameter corresponding to the immediate facial image of described illumination value.
Preferably, describedly in the face database prestored, search facial image immediate with described illumination value, and obtain pretreatment parameter corresponding to the immediate facial image of described illumination value, comprising:
Set up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
According to described discrete equation, calculate immediate with reference to illumination value with the illumination value of the facial image collected, and obtain the described pretreatment parameter with reference to illumination value.
Particularly, if L is the set of the illumination value of face images in face database, then discrete equation F one to one can be created, as formula (10).
γ
i=F(l
i),i=1,...,|L| (10)
Wherein, l
ithe illumination value of i-th facial image in L, γ
ithe precalculated gamma parameter γ of i-th facial image.Suppose that the illumination value of the facial image collected is l, then can by calculating gamma parameter γ
*process calculate l
*, as shown in Equation (11).
So, by formula (11), gamma parameter γ
*the time complexity of optimizing process becomes illumination value l
*the time complexity of optimizing process, as formula (12).
T
1’=n·O((l
i-l)
2)+O(n)+O(1)=O(n) (12)
In sum, the All Time complexity of the light irradiation preprocess method of embodiment of the present invention facial image obtains by formula (13).
T'=T
1'+T
2=O(n)+O(H·W) (13)
From formula (13), All Time complexity is a fixing value, only depends on the size of face database and often opens the size of facial image.GIC algorithm before comparing, the light irradiation preprocess method of embodiment of the present invention facial image can greatly reduce time complexity and calculate the resource consumed.
Step 103: illumination pretreatment is carried out to the facial image collected according to described pretreatment parameter.
Preferably, described method also comprises: carry out recognition of face process to the facial image after illumination pretreatment.
The embodiment of the present invention also describes a kind of illumination pretreatment device of facial image, and as shown in Figure 2, described device comprises: collecting unit 21, acquiring unit 22, search unit 23 and processing unit 24; Wherein,
Described collecting unit 21, during for gathering facial image, triggers described acquiring unit 22;
Described acquiring unit 22, for receive described collecting unit 21 triggering after, obtain the illumination value of environment residing for terminal;
Describedly searching unit 23, for searching facial image immediate with described illumination value in the face database prestored, and obtaining pretreatment parameter corresponding to the immediate facial image of described illumination value;
Described processing unit 24, for carrying out illumination pretreatment according to described pretreatment parameter to the facial image collected.
Preferably, on the basis of the illumination pretreatment device of the facial image shown in Fig. 2, the illumination pretreatment device of the facial image of the embodiment of the present invention also comprises: pretreatment unit (not shown in Fig. 2) and human face data library unit (not shown in Fig. 2); Wherein,
Described collecting unit 21, also for gathering facial image during multiple intensity of illumination respectively, and obtains the facial image of several different illumination intensity;
Described acquiring unit 22, also for obtaining the illumination value of the facial image of several different illumination intensity described, and as reference illumination value;
Described pretreatment unit, for calculating the pretreatment parameter of the facial image of several different illumination intensity described respectively according to illumination pretreatment algorithm;
Described human face data library unit, for the pretreatment parameter of the illumination value of the facial image of the facial image according to several different illumination intensity described, several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
Preferably, when described illumination pretreatment algorithm is GIC algorithm, described pretreatment parameter is gamma parameter.
Preferably, as shown in Figure 3, search unit 23 described in comprise: set up subelement 231 and computation subunit 232; Wherein,
Describedly set up subelement 231, for setting up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
Described computation subunit 232, for according to described discrete equation, calculates immediate with reference to illumination value with the illumination value of the facial image collected, and obtains the described pretreatment parameter with reference to illumination value.
Preferably, on the basis of the illumination pretreatment device of the facial image shown in Fig. 2, the illumination pretreatment device of the facial image of the embodiment of the present invention also comprises recognition unit (not shown in Fig. 2), for carrying out recognition of face process to the facial image after illumination pretreatment.
It will be appreciated by those skilled in the art that the practical function of each unit in the illumination pretreatment device of the facial image shown in Fig. 2, Fig. 3 and subelement thereof can refer to the associated description of the light irradiation preprocess method of aforementioned facial image and understands.Each unit in the illumination pretreatment device of the facial image shown in Fig. 2, Fig. 3 and the function of subelement thereof realize by the program run on processor, also realize by concrete logical circuit.
In actual applications, the collecting unit 21 in the illumination pretreatment device of above-mentioned facial image can be realized by camera head; Acquiring unit 22 can be realized by the light sensor of terminal built-in; Search unit 23 and processing unit 24 can by central processing unit (CPU, Central Processing Unit) or digital signal processor (DSP, Digital Signal Processor) or programmable logic array (FPGA, Field-Programmable Gate Array) realization.
The embodiment of the present invention also describes a kind of terminal, and as shown in Figure 4, described terminal comprises the illumination pretreatment device of the facial image shown in Fig. 2, comprising: collecting unit 21, acquiring unit 22, search unit 23 and processing unit 24; Wherein,
Described collecting unit 21, during for gathering facial image, triggers described acquiring unit;
Described acquiring unit 22, for receive described collecting unit triggering after, obtain the illumination value of environment residing for terminal;
Describedly searching unit 23, for searching facial image immediate with described illumination value in the face database prestored, and obtaining pretreatment parameter corresponding to the immediate facial image of described illumination value;
Described processing unit 24, for carrying out illumination pretreatment according to described pretreatment parameter to the facial image collected.
Preferably, on the basis of the illumination pretreatment device of the facial image shown in Fig. 2, the illumination pretreatment device of the facial image of the embodiment of the present invention also comprises: pretreatment unit (not shown in Fig. 2) and human face data library unit (not shown in Fig. 2); Wherein,
Described collecting unit 21, also for gathering facial image during multiple intensity of illumination respectively, and obtains the facial image of several different illumination intensity;
Described acquiring unit 22, also for obtaining the illumination value of the facial image of several different illumination intensity described, and as reference illumination value;
Described pretreatment unit, for calculating the pretreatment parameter of the facial image of several different illumination intensity described respectively according to illumination pretreatment algorithm;
Described human face data library unit, for the pretreatment parameter of the illumination value of the facial image of the facial image according to several different illumination intensity described, several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
Preferably, when described illumination pretreatment algorithm is GIC algorithm, described pretreatment parameter is gamma parameter.
Preferably, as shown in Figure 3, search unit 23 described in comprise: set up subelement 231 and computation subunit 232; Wherein,
Describedly set up subelement 231, for setting up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
Described computation subunit 232, for according to described discrete equation, calculates immediate with reference to illumination value with the illumination value of the facial image collected, and obtains the described pretreatment parameter with reference to illumination value.
Preferably, on the basis of the illumination pretreatment device of the facial image shown in Fig. 2, the illumination pretreatment device of the facial image of the embodiment of the present invention also comprises recognition unit (not shown in Fig. 2), for carrying out recognition of face process to the facial image after illumination pretreatment.
It will be appreciated by those skilled in the art that the function that the illumination pretreatment device of the facial image in the terminal shown in Fig. 4 can refer to each unit in the illumination pretreatment device of the facial image shown in Fig. 2, Fig. 3 and subelement thereof is understood.
In actual applications, the collecting unit 21 in the illumination pretreatment device of the facial image in the terminal shown in Fig. 4 can be realized by camera head; Acquiring unit 22 can be realized by the light sensor of terminal built-in; Search unit 23 and processing unit 24 can by central processing unit (CPU, Central Processing Unit) or digital signal processor (DSP, Digital Signal Processor) or programmable logic array (FPGA, Field-Programmable Gate Array) realization.
Be described in further detail below in conjunction with the effect of specific embodiment to the light irradiation preprocess method of facial image provided by the invention.
Have chosen HUN-MFD standard in the present embodiment and move relevant face image data in face database, one has 349 facial images, 12 samples, comprises 4 women and 8 male sex.The facial image quantity of each sample from 14 to 55 not etc., contains the facial image under multiple scene.Under the facial image of each sample all chooses frontal pose, different illumination conditions, do not have to make up and the situation of expression.Training set, is the face database in the embodiment of the present invention, contains 246 facial images, and 103 remaining facial images are as test set.Repeater's face image is not had between training set and test set.Each is opened, and photo is all prior is normalized to Gray Face image through cutting, and the facial image of unified employing 128 × 128 pixel.Fig. 5 shows the facial image under the different illumination conditions of three samples.
Have employed three kinds of classical face characteristics in an embodiment and carry out assessed for performance as a comparison: PCA [11], LDA [12] and LBP [13].The most contiguous (NN) method is for calculating similarity.Here all people's face discrimination all refers to Rank-1.The parameter of three-type-person's face feature is described below respectively:
PCA: have employed the major component that face characteristic length is 50, each feature saves the quantity of information of in former figure 92%;
LDA: owing to there being 13 samples, adopt default value, therefore characteristic length is 12;
LBP: as a kind of local feature, the radius of LBP feature is set to 1, and around each pixel, 8 pixels can be used to participate in calculating.What give tacit consent to is 8 × 8 by facial image piecemeal.
In this performance testing example, setting up two running environment and contrast, is computerized environment and mobile environment respectively.Which environment is embodiment on computerized environment mainly in order to verify that whether mobile GIC algorithm can produce a desired effect under computer and mobile two environment, and compare and be more suitable for.The priority of program is all set to limit priority at utmost to prevent the impact of program on embodiment of other parallel runnings by the embodiment in two environment.The design parameter of two embodiment environment is as follows:
Computer run environment: the application program of the embodiment of the present invention operates in Windows PC, and its CPU is
dual-Core E5200@2.50GHz, internal memory 2G, operating system is Windows 7 Ultimate Service Pack 1.The dependence storehouse of using during program development includes built-in function in Visual Studio 2010 and OpenCV 2.4.2;
Mobile execution environment: embodiment program operates on Android cell phone, its model is Samsungi9250 (Google Galaxy Nexus), CPU is OMAP4460 Dual-Core 1228 MHz, internal memory 1G, operating system is Android 4.1.2(Jelly Bean), operating system nucleus version is Linux3.0.31-g4f6d371.The dependence storehouse of using during program development includes android-ndk-r8b and OpenCV2.4.2.In order to keep the consistance between the program of computer environment, under mobile environment is transplanted to by identical C++ code, and do not write the program being rewritten into Java version from C++ in addition.Therefore in order to meet such demand, application program have employed to develop based on Java local interface (JNI, Java Native Interface) technology.
When running GIC algorithm, usually empirically the scope of γ is arranged to (0,5].Fig. 6 respectively illustrates the facial image after for the original facial image under a sample different illumination conditions, GIC and mobile GIC.
Table 1 shows mobile GIC algorithm (light irradiation preprocess method for the facial image in the embodiment of the present invention) and compares with GIC algorithm, the speed-up ratio of preprocessing part when using different step-length respectively under computer environment and mobile environment.
Table 1
Table 2 shows the Rank-1 face identification rate used respectively on mobile phone under PCA, LDA and LBP face characteristic and different pretreatments method.
Table 2
Under table 3 shows mobile environment, comprise the speed-up ratio between the mobile Gamma intensity rectification of the complete face recognition process of preprocessing part with the rectification of Gamma intensity.
Table 3
Can be drawn by the data of statistics in upper table 1 to table 3:
The acceleration effect of mobile GIC algorithm is in a mobile environment than good under computer environment;
With do not do compared with compensation, mobile GIC algorithm greatly improves face identification rate, but compared with GIC algorithm, in the algorithm situation of PCA and LDA, the discrimination of mobile GIC algorithm is lower slightly, but during with local feature algorithm LBP, mobile GIC algorithm has higher discrimination;
Due to computing cost extremely low in a mobile environment and power consumption, mobile GIC algorithm is particularly suitable for running on the mobile apparatus, especially when using LDA, both improving discrimination, and significantly reducing time complexity again.
The above, be only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.
Claims (11)
1. a light irradiation preprocess method for facial image, is characterized in that, described method comprises:
When gathering facial image, and obtain the illumination value of environment residing for terminal;
In the face database prestored, search facial image immediate with described illumination value, and obtain pretreatment parameter corresponding to the immediate facial image of described illumination value;
According to described pretreatment parameter, illumination pretreatment is carried out to the facial image collected.
2. method according to claim 1, is characterized in that, described method also comprises:
Gather facial image during multiple intensity of illumination respectively, and obtain the facial image of several different illumination intensity;
The illumination value of the facial image of several different illumination intensity described is obtained by described light sensor, and as reference illumination value;
The pretreatment parameter of the facial image of several different illumination intensity described is calculated respectively according to illumination pretreatment algorithm;
According to the pretreatment parameter of the facial image of several different illumination intensity described, the illumination value of the facial image of several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
3. method according to claim 1, is characterized in that, when described illumination pretreatment algorithm is gamma intensity rectification GIC algorithm, described pretreatment parameter is gamma parameter.
4. the method according to any one of claims 1 to 3, is characterized in that, describedly in the face database prestored, searches facial image immediate with described illumination value, and obtains pretreatment parameter corresponding to the immediate facial image of described illumination value, comprising:
Set up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
According to described discrete equation, calculate immediate with reference to illumination value with the illumination value of the facial image collected, and obtain the described pretreatment parameter with reference to illumination value.
5. the method according to any one of claims 1 to 3, is characterized in that, described method also comprises:
Recognition of face process is carried out to the facial image after illumination pretreatment.
6. an illumination pretreatment device for facial image, is characterized in that, described device comprises: collecting unit, acquiring unit, search unit and processing unit; Wherein,
Described collecting unit, during for gathering facial image, triggers described acquiring unit;
Described acquiring unit, for receive described collecting unit triggering after, obtain the illumination value of environment residing for terminal;
Describedly searching unit, for searching facial image immediate with described illumination value in the face database prestored, and obtaining pretreatment parameter corresponding to the immediate facial image of described illumination value;
Described processing unit, for carrying out illumination pretreatment according to described pretreatment parameter to the facial image collected.
7. device according to claim 6, is characterized in that, described device also comprises pretreatment unit and human face data library unit;
Described collecting unit, also for gathering facial image during multiple intensity of illumination respectively, and obtains the facial image of several different illumination intensity;
Described acquiring unit, also for obtaining the illumination value of the facial image of several different illumination intensity described, and as reference illumination value;
Described pretreatment unit, for calculating the pretreatment parameter of the facial image of several different illumination intensity described respectively according to illumination pretreatment algorithm;
Described human face data library unit, for the pretreatment parameter of the illumination value of the facial image of the facial image according to several different illumination intensity described, several different illumination intensity described and the facial image of several different illumination intensity described, set up face database.
8. device according to claim 6, is characterized in that, when described illumination pretreatment algorithm is GIC algorithm, described pretreatment parameter is gamma parameter.
9. the device according to any one of claim 6 to 8, is characterized in that, described in search unit and comprise: set up subelement and computation subunit; Wherein,
Describedly set up subelement, for setting up the discrete equation between the illumination value of facial image of several different illumination intensity described and the pretreatment parameter of the facial image of several different illumination intensity described;
Described computation subunit, for according to described discrete equation, calculates immediate with reference to illumination value with the illumination value of the facial image collected, and obtains the described pretreatment parameter with reference to illumination value.
10. the device according to any one of claim 6 to 8, is characterized in that, described device also comprises recognition unit, for carrying out recognition of face process to the facial image after illumination pretreatment.
11. 1 kinds of terminals, is characterized in that, described terminal comprises the illumination pretreatment device of the facial image described in any one of claim 6 to 10.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310366699.9A CN104424483A (en) | 2013-08-21 | 2013-08-21 | Face image illumination preprocessing method, face image illumination preprocessing device and terminal |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310366699.9A CN104424483A (en) | 2013-08-21 | 2013-08-21 | Face image illumination preprocessing method, face image illumination preprocessing device and terminal |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104424483A true CN104424483A (en) | 2015-03-18 |
Family
ID=52973396
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310366699.9A Pending CN104424483A (en) | 2013-08-21 | 2013-08-21 | Face image illumination preprocessing method, face image illumination preprocessing device and terminal |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104424483A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106295596A (en) * | 2016-08-17 | 2017-01-04 | 深圳市金立通信设备有限公司 | A kind of unlocking method based on recognition of face and terminal |
CN106295571A (en) * | 2016-08-11 | 2017-01-04 | 深圳市赛为智能股份有限公司 | The face identification method of illumination adaptive and system |
CN106534576A (en) * | 2016-12-07 | 2017-03-22 | 深圳市传奇数码有限公司 | Side face unlocking method applied to mobile phone |
CN107705245A (en) * | 2017-10-13 | 2018-02-16 | 北京小米移动软件有限公司 | Image processing method and device |
CN107944420A (en) * | 2017-12-07 | 2018-04-20 | 北京旷视科技有限公司 | The photo-irradiation treatment method and apparatus of facial image |
CN108537749A (en) * | 2018-03-29 | 2018-09-14 | 广东欧珀移动通信有限公司 | Image processing method, device, mobile terminal and computer readable storage medium |
CN109325448A (en) * | 2018-09-21 | 2019-02-12 | 广州广电卓识智能科技有限公司 | Face identification method, device and computer equipment |
CN110457976A (en) * | 2018-05-08 | 2019-11-15 | 上海箩箕技术有限公司 | Fingerprint imaging method and fingerprint imaging system |
CN112651993A (en) * | 2020-11-18 | 2021-04-13 | 合肥市卓迩无人机科技服务有限责任公司 | Moving target analysis and recognition algorithm for multi-path 4K quasi-real-time spliced video |
WO2021208373A1 (en) * | 2020-04-14 | 2021-10-21 | 北京迈格威科技有限公司 | Image identification method and apparatus, and electronic device and computer-readable storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101046847A (en) * | 2007-04-29 | 2007-10-03 | 中山大学 | Human face light alignment method based on secondary multiple light mould |
CN101201900A (en) * | 2007-11-06 | 2008-06-18 | 重庆大学 | Method for regulating human face image illumination based on multilevel wavelet disintegrating and spline interpolation |
CN101694691A (en) * | 2009-07-07 | 2010-04-14 | 北京中星微电子有限公司 | Method and device for synthesizing facial images |
US20100195912A1 (en) * | 2009-02-05 | 2010-08-05 | Naohisa Nakada | Imaging device, image composition and display device, and image composition method |
CN101916384A (en) * | 2010-09-01 | 2010-12-15 | 汉王科技股份有限公司 | Facial image reconstruction method and device and face recognition system |
CN102682276A (en) * | 2011-12-20 | 2012-09-19 | 河南科技大学 | Face recognition method and base image synthesis method under illumination change conditions |
-
2013
- 2013-08-21 CN CN201310366699.9A patent/CN104424483A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101046847A (en) * | 2007-04-29 | 2007-10-03 | 中山大学 | Human face light alignment method based on secondary multiple light mould |
CN101201900A (en) * | 2007-11-06 | 2008-06-18 | 重庆大学 | Method for regulating human face image illumination based on multilevel wavelet disintegrating and spline interpolation |
US20100195912A1 (en) * | 2009-02-05 | 2010-08-05 | Naohisa Nakada | Imaging device, image composition and display device, and image composition method |
CN101694691A (en) * | 2009-07-07 | 2010-04-14 | 北京中星微电子有限公司 | Method and device for synthesizing facial images |
CN101916384A (en) * | 2010-09-01 | 2010-12-15 | 汉王科技股份有限公司 | Facial image reconstruction method and device and face recognition system |
CN102682276A (en) * | 2011-12-20 | 2012-09-19 | 河南科技大学 | Face recognition method and base image synthesis method under illumination change conditions |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106295571A (en) * | 2016-08-11 | 2017-01-04 | 深圳市赛为智能股份有限公司 | The face identification method of illumination adaptive and system |
CN106295596A (en) * | 2016-08-17 | 2017-01-04 | 深圳市金立通信设备有限公司 | A kind of unlocking method based on recognition of face and terminal |
CN106534576A (en) * | 2016-12-07 | 2017-03-22 | 深圳市传奇数码有限公司 | Side face unlocking method applied to mobile phone |
CN107705245A (en) * | 2017-10-13 | 2018-02-16 | 北京小米移动软件有限公司 | Image processing method and device |
CN107944420A (en) * | 2017-12-07 | 2018-04-20 | 北京旷视科技有限公司 | The photo-irradiation treatment method and apparatus of facial image |
CN107944420B (en) * | 2017-12-07 | 2020-10-27 | 北京旷视科技有限公司 | Illumination processing method and device for face image |
CN108537749A (en) * | 2018-03-29 | 2018-09-14 | 广东欧珀移动通信有限公司 | Image processing method, device, mobile terminal and computer readable storage medium |
CN110457976A (en) * | 2018-05-08 | 2019-11-15 | 上海箩箕技术有限公司 | Fingerprint imaging method and fingerprint imaging system |
CN109325448A (en) * | 2018-09-21 | 2019-02-12 | 广州广电卓识智能科技有限公司 | Face identification method, device and computer equipment |
WO2021208373A1 (en) * | 2020-04-14 | 2021-10-21 | 北京迈格威科技有限公司 | Image identification method and apparatus, and electronic device and computer-readable storage medium |
CN112651993A (en) * | 2020-11-18 | 2021-04-13 | 合肥市卓迩无人机科技服务有限责任公司 | Moving target analysis and recognition algorithm for multi-path 4K quasi-real-time spliced video |
CN112651993B (en) * | 2020-11-18 | 2022-12-16 | 合肥市卓迩无人机科技服务有限责任公司 | Moving target analysis and recognition algorithm for multi-path 4K quasi-real-time spliced video |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104424483A (en) | Face image illumination preprocessing method, face image illumination preprocessing device and terminal | |
US20210012198A1 (en) | Method for training deep neural network and apparatus | |
CN109146892B (en) | Image clipping method and device based on aesthetics | |
WO2021073418A1 (en) | Face recognition method and apparatus, device, and storage medium | |
WO2022199583A1 (en) | Image processing method and apparatus, computer device, and storage medium | |
CN110084299B (en) | Target detection method and device based on multi-head fusion attention | |
CN108875602A (en) | Monitor the face identification method based on deep learning under environment | |
CN107392214B (en) | Target detection method based on full-volume integral crack network | |
CN106295613A (en) | A kind of unmanned plane target localization method and system | |
CN104778481A (en) | Method and device for creating sample library for large-scale face mode analysis | |
WO2021027692A1 (en) | Visual feature library construction method and apparatus, visual positioning method and apparatus, and storage medium | |
CN112164005B (en) | Image color correction method, device, equipment and storage medium | |
CN111695463B (en) | Training method of face impurity detection model and face impurity detection method | |
CN113704531A (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
CN111209907A (en) | Artificial intelligent identification method for product characteristic image in complex light pollution environment | |
JP2023527615A (en) | Target object detection model training method, target object detection method, device, electronic device, storage medium and computer program | |
CN113487610B (en) | Herpes image recognition method and device, computer equipment and storage medium | |
CN109543744B (en) | Multi-category deep learning image identification method based on Loongson group and application thereof | |
CN112465709B (en) | Image enhancement method, device, storage medium and equipment | |
CN112001399A (en) | Image scene classification method and device based on local feature saliency | |
CN110232131A (en) | Intention material searching method and device based on intention label | |
CN112668509B (en) | Training method and recognition method of social relation recognition model and related equipment | |
Chen et al. | An improved dense-to-sparse cross-modal fusion network for 3D object detection in RGB-D images | |
CN116912774A (en) | Infrared image target identification method, electronic device and storage medium of power transmission and transformation equipment based on edge calculation | |
CN106469437B (en) | Image processing method and image processing apparatus |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150318 |
|
RJ01 | Rejection of invention patent application after publication |