CN101241550B - Iris image quality judgment method - Google Patents

Abstract

The invention provides a method for judging quality of iris image, which comprises extracting original normalized iris image by positioning and normalization, performing 4-layer multi-resolution decomposition on the normalized iris image, and counting resolution 2^-1The number of points with larger amplitude of detail components is compared with a preset threshold value, whether the image has eyelid and eyelash occlusion problems can be judged, and then the resolution 2 is calculated^-2The variance of the detail component is compared with a preset threshold value, and whether the image definition meets the requirements of the system can be judged. The iris image quality judging method has stable performance in different illumination environments, and can quickly and accurately judge the iris image quality.

Description

Iris image quality judgment method

[ technical field ] A method for producing a semiconductor device

The invention belongs to the technical field of image processing, and mainly relates to an iris identity recognition technology in biological characteristic identification.

[ background of the invention ]

As electronic devices such as computers, ATMs, mobile phones, door access control systems, and the like continuously enter our daily lives, personal security and convenient identity authentication technologies become more and more urgent. However, existing systems based on smart cards, identification numbers and passwords can only wander between security and convenience, sufficient security has never been achieved, and better security comes with inconvenience. To achieve higher security, we must use a more complex and less convenient password because if we use the same password for different machines around us, we get convenience while increasing security concerns. For this reason, many safe and convenient solutions have been sought. The biological recognition technology provides a completely new field for the biological recognition technology. Biometric identification technology is the identification and verification of a person using physiological or behavioral characteristics of the person. These physiological characteristics include fingerprints, palm prints, voice, signatures, etc. to identify an individual. The physiological characteristics are not forgotten like passwords or lost like keys, so the method is considered to be a more reliable personal identity authentication method. The iris identity recognition is a new biological recognition technology, and the iris is used as the basis of the identity recognition, so that the iris identity recognition has the advantages of high uniqueness, high stability, natural anti-counterfeiting property, no invasion and the like. For details, see the literature: jain K, Aren Ross, Salil Prabhakar, "An Introduction to biological Recognition", IEEE Transaction on Circuits and Systems for Video Technology, Volume 14, No.1, pp.4-20, 2004 and literature: john G.Daugman, "How is Recognition Works," IEEE Transaction on Circuits and Systems for Video Technology, Volume 14, Issue 1, pp.21-30, 2004.

The quality evaluation of the iris image is an important part in the whole automatic iris recognition technology, and ensures that the processed iris image meets the quality requirement of the system. Therefore, false recognition and false recognition caused by the quality problem of the iris image are avoided. In practice, due to the focal length problem of the acquisition equipment during shooting, the rotation problem of eyeballs at the moment of shooting, and the partial shielding of eyelids and eyelashes on the iris, the acquired iris image cannot be subjected to subsequent feature extraction. At present, an effective iris image quality evaluation model is not provided in the existing algorithm, so that a set of general and feasible evaluation models is established, and the details are shown in the literature: chen Ji, Hu Guingshu, "Iris Image Quality Evaluation based on Wavelet Packet Decomposition", Journal of Tsinghua University (Sci & Tech), Volume 43, No.3, pp.377-380, 2003 and literature: li Ma, Tieniu Tan, Yunhong Wang, Dexin Zhang, "effective Iris registration by chromatography keying Local values," IEEE Transaction on Image Processing, Volume 13, No.6, pp.739-750, 2004.

The existing iris quality judgment methods comprise:

(1) a fast fourier transform based method. The method carries out two-dimensional fast Fourier transform on pixel points in two rectangular blocks on an iris area, and then analyzes whether an image is clear or not and eyelashes are shielded or not through statistics of high-frequency, medium-frequency and low-frequency energy of the pixel points. The model is not strong in universality, and clear iris images with few textures are easily judged to be low-quality iris images in an error mode. For details, see the literature: li Ma, Tieniu Tan, Yunhong Wang, Dexin Zhang, "Personal Identification based on Iris Texture Analysis," IEEE Transactions on Pattern Analysis and Machine Analysis, Volume 25, No.12, pp.1519-1533.

(2) A method based on wavelet packet decomposition. The method selects the sub-band with the most concentrated texture high-frequency component distribution as a characteristic sub-band, and takes the energy of the sub-band as a criterion for judging the image quality. The method has a disadvantage that it cannot judge an iris image having a problem due to occlusion of eyelashes. For details, see the literature: chen Ji, Hu Guingshu, "Iris Image Quality Evaluation based on Wavelet Packet Decomposition," Journal of Tsinghua University (Sci & Tech), Volume 43, No.3, pp.377-380, 2003.

(3) Method based on image sharpness, internal and external eccentricity and iris visibility. The method establishes three indexes of image quality such as image definition, internal and external eccentricity and iris visibility, and meets the requirement of real-time quality evaluation on the iris image. The method has the disadvantages of sensitivity to illumination conditions and poor stability. Xing Lei, Shi Pengfei, "A Quality Evaluation Method of Iris Image", Chinese Journal of Stereology and Image Analysis, volume.8, No.2, pp.108-113, 2003.

The iris image quality judgment algorithm has problems to a certain extent, and is too large in calculation amount, sensitive to illumination, low in universality and the like.

[ summary of the invention ]

The invention aims to establish an iris image quality judgment method with stronger universality, which can accurately detect iris images shielded by eyelash and iris images with insufficient definition and make the illumination condition range applicable to the algorithm wider.

The purpose of the invention is realized as follows:

1. a method for judging quality of an iris image is characterized by comprising the following steps:

step 1, carrying out image acquisition on an iris in human eyes through a camera device, and obtaining a normalized iris image f (x, y) with the size of M multiplied by N from an original gray image containing the iris image; (x, y) represents the coordinates of the pixel points, and f (x, y) represents the gray value of the pixel points with the coordinates of (x, y);

step 2, performing 4-layer two-dimensional wavelet transform on the normalized iris image obtained in the step 1; specifically, the formula of the two-dimensional wavelet transform is:

and

wherein,

is a resolution of 2^jThe scale factor of (a) is lower,

is a resolution of 2^jThe value range of the 4-layer wavelet transformation j is { -1, -2, -3, -4}, i { -H, V, D }, and details in the horizontal direction, the vertical direction and the diagonal direction are added;

in order to be a function of the scale,

selecting a wavelet as a wavelet function, wherein the selected wavelet is a DMEyer wavelet; scale function for performing two-dimensional wavelet transform

Is a function of two one-dimensional scalesAndthe product of (a); horizontally sensitive wavelet function for two-dimensional wavelet transform

As a one-dimensional wavelet function psi_j，m(x) And a one-dimensional scale function

The product of (a); vertical sensitive wavelet function for two-dimensional wavelet transformAs a function of a one-dimensional scale

And a one-dimensional wavelet function psi_j，n(y) the product of (a); diagonally sensitive wavelet function for two-dimensional wavelet transform

For two one-dimensional wavelet functions psi_j，m(x) And psi_j，n(y) the product of (a);

step 3, obtaining resolution 2 through step 2^-1Lower horizontal direction wavelet coefficient

Vertical direction wavelet coefficient

And diagonal direction wavelet coefficients

Reconstructing the original normalized Iris image at resolution 2^-1Details of

I.e. high frequency detail components; specifically, the reconstruction formula is:

wherein

Representation resolution 2^jThe value of the detail component of the lower coordinate (x, y),

representation resolution 2^jThe wavelet coefficients of the lower one(s),

is a two-dimensional wavelet sensitive to the i direction; sigma is an accumulation operator;

step 4, calculating the high-frequency detail component obtained in the step 3

Absolute value of (2)

In particular, if

If it is not

$\left|D_{2^{-1}}f(x,y)\right|=-D_{2^{-1}}f(x,y);$

Step 5, counting the obtained in step 4

Is greater than a threshold value V_oNumber of pixel points of will

Is greater than a threshold value V_oAs boundary points of eyelids and eyelashes; v_oIn order to determine whether the pixel point is the threshold of the boundary point of the eyelids and eyelashes, the specific calculation formula is as follows:where M and N are the width and height of the original normalized iris image,

step 6, comparing the values calculated in step 5

Value and threshold T_NComparing; if it is not

The image is considered as an image of the eyelid and eyelash occlusion if

The image is considered to be a processable normal iris image;

step 7, according to the method in step 3, passing the resolution 2 obtained in step 2^-2Lower horizontal direction wavelet coefficient

Vertical direction wavelet coefficient

And diagonal direction wavelet coefficientsReconstructing the original normalized Iris image at resolution 2^-2Details of

Step 8, calculating the result obtained in step 7

The variance of (a); the specific calculation formula is as follows:wherein Var is resolution 2^-2The variance of the detail components, M and N, are the width and height of the original normalized iris image, respectively, and the calculation formula of M is: <math><mrow><mi>m</mi><mo>=</mo><munderover><mi>&Sigma;</mi><mrow><mi>x</mi><mo>=</mo><mn>1</mn></mrow><mi>M</mi></munderover><munderover><mi>&Sigma;</mi><mrow><mi>y</mi><mo>=</mo><mn>1</mn></mrow><mi>N</mi></munderover><msub><mi>D</mi><mrow><msup><mn>2</mn><mrow><mo>-</mo><mn>2</mn></mrow></msup><mi>f</mi><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow></mrow></msub><mo>/</mo><mi>M</mi><mo>&times;</mo><mi>N</mi><mo>;</mo></mrow></math>

step 9, the variance Var obtained in the step 8 and a threshold value T for judging the definition of the iris_vComparing, if Var is more than or equal to T_vIf so, the definition of the iris image meets the system requirement; if Var < T_vAnd the definition of the iris image does not meet the system requirement.

The method for judging the quality of the iris image is characterized in that the threshold T mentioned in the step 6_NIs an image for judging whether the image is a processable maskless eyelid and eyelash occlusion image, T_NIs associated with a subsequent matching algorithm, which is set to

Matching iris images of (1), T corresponding to the lowest error rate_N。

The invention adopts a multi-resolution analysis method, and correctly evaluates the quality of the iris image by analyzing detail components under different resolutions. The method comprises the steps of firstly positioning an original iris image to obtain a normalized iris image. By performing 4-layer wavelet transform on the normalized iris image, the statistical resolution is 2^-1And comparing the number of the larger amplitude points of the detail components with a preset threshold value, and judging whether the iris image has the problem of eyelid and eyelash occlusion. Thereafter, the resolution 2 is counted^-2And comparing the variance of the detail components with a preset threshold value to judge whether the iris image is clear. The method for analyzing the quality of the iris image by utilizing the multi-resolution thought is a characteristic of the method, and compared with a general iris image quality evaluation method, the method has strong universality and stability and is not easily influenced by illumination.

[ description of the drawings ]

FIG. 1 is an original image containing an iris;

wherein 1 represents a pupil; 2 represents an iris; 3 denotes the spot in the pupil; 4 denotes the inner edge of the iris; and 5 denotes the outer edge of the iris.

FIG. 2 is a positioning result graph and a normalization graph;

wherein, (a) is a positioning result graph; (b) the iris normalized image is obtained.

FIG. 3 is an original normalized iris image and its high frequency detail components;

wherein (a) is a normalized iris image with eyelid occlusion, (b) is a high frequency detail component of (a), (c) is a normalized iris image with eyelash occlusion, and (d) is a high frequency detail component of (c).

[ detailed description ] embodiments

For convenience in describing the present disclosure, certain terms are first defined.

Definition 1: the iris. The center of the eyeball is a black pupil, and the annular tissue between the outer edges of the pupil is the iris. It exhibits interlaced textural features like spots, filaments, stripes, crypts. The irises of the same person hardly change during the life of the person, and the irises of different persons are completely different.

Definition 2: a grayscale image. An image that contains only luminance information and no other color information in the image.

Definition 3: and normalizing the iris image. After the original iris image is positioned, normalization operation is carried out to eliminate the problems of head rotation, inconsistent shooting distance, pupil zooming and the like during shooting to obtain an image, and the normalized iris image has the same size.

Definition 4: and (5) performing wavelet transformation. The local analysis method of time (space) frequency carries out multi-scale thinning on the signal (function) step by step through telescopic translation operation, and can focus on any details of the signal.

Definition 5: a scale factor. In performing wavelet transform, the coefficients obtained after convolution of the original signal with a scale function are used to reconstruct the approximate components of the signal. For two-dimensional wavelet transform, the specific calculation formula of the scale coefficient is as follows:

wherein f (x, y) is the original signal,in order to be a function of the scale,

is a scale factor.

Definition 6: a scale function. The scale function is a binary scale, square integrable function consisting of integer translation and real number

Set of constituent spreading functions, i.e. sets

Wherein

j, m ∈ Z. Defining the Scale function of the two-dimensional wavelet transform in 5

Is a function of two one-dimensional scales

And

the product of (a).

Definition 7: wavelet coefficients. In the wavelet transform, the coefficients obtained after the convolution of the original signal with the wavelet function are used to reconstruct the detail components of the signal. For two-dimensional wavelet transform, the specific calculation formula of the wavelet coefficient is as follows:

wherein f (x, y) is the original signal,

in order to be a function of the wavelet,

are wavelet coefficients.

Definition 8: a wavelet function. The wavelet function is used for describing the difference between two adjacent scale spaces and is an expansion function set consisting of psi (x), i.e. the set psi_j，k(x) And (4) dividing. Wherein psi_j，m(x)＝2^j/2ψ(2^jx-m), j, m ∈ Z. Defining a horizontally sensitive wavelet function of the two-dimensional wavelet transform in 5

As a one-dimensional wavelet function psi_j，m(x) And a one-dimensional scale functionThe product of (a); vertical sensitive wavelet function for two-dimensional wavelet transform

As a function of a one-dimensional scale

And a one-dimensional wavelet function psi_j，n(y) the product of (a); diagonally sensitive wavelet function for two-dimensional wavelet transformFor two one-dimensional wavelet functions psi_j，m(x) And psi_j，nThe product of (y).

Definition 9: a DMeyer wavelet. The discrete form of the Meyer wavelet is an effective approximation of the Meyer wavelet, and can be regarded as a discretized Meyer wavelet, with bi-orthogonality. It not only keeps the good frequency division characteristic of Meyer wavelet, but also can raise the speed of numerical calculation.

Definition 10: and (4) multi-resolution analysis. The idea of multiresolution analysis is mainly to consider the original image f (x, y) as having a resolution of 2⁰The approximation is further decomposed into a coarse resolution 2, which is an approximation under 1^JApproximate component at (J < 0) and a series of high resolution 2^j(J > J) the sum of the detail components is asymptotically approximated.

Definition 11: a detail component. Any one image can be decomposed into main body information and detail texture information, and according to the idea of multi-resolution analysis, the detail components refer to the detail texture information in different frequency band ranges.

Definition 12: and (6) matching. A particular thing is correctly classified into a certain category.

Definition 13. The rate of false identification. The probability of categorizing things of the same class into other classes is commonly expressed in FMR.

The iris image quality judging method according to the present invention comprises the steps of:

1. a method for judging quality of an iris image is characterized by comprising the following steps:

step 1, carrying out image acquisition on an iris in human eyes through a camera device, and obtaining a normalized iris image f (x, y) with the size of M multiplied by N from an original gray image containing the iris image; (x, y) represents the coordinates of the pixel points, and f (x, y) represents the gray value of the pixel points with the coordinates of (x, y);

step 2, performing 4-layer two-dimensional wavelet transform on the normalized iris image obtained in the step 1; specifically, the formula of the two-dimensional wavelet transform is:andwherein,

is a resolution of 2^jThe scale factor of (a) is lower,

is a resolution of 2^jThe value range of the 4-layer wavelet transformation j is { -1, -2, -3, -4}, i { -H, V, D }, and details in the horizontal direction, the vertical direction and the diagonal direction are added;

in order to be a function of the scale,

selecting a wavelet as a wavelet function, wherein the selected wavelet is a DMEyer wavelet;

step 3, obtaining resolution 2 through step 2^-1Down horizontal directionWavelet coefficient

Vertical direction wavelet coefficient

And diagonal direction wavelet coefficientsReconstructing the original normalized Iris image at resolution 2^-1Details of

I.e. high frequency detail components; specifically, the reconstruction formula is:

wherein

Representation resolution 2^jThe value of the detail component of the lower coordinate (x, y),

representation resolution 2^jThe wavelet coefficients of the lower one(s),

is a two-dimensional wavelet sensitive to the i direction; sigma is an accumulation operator;

step 4, calculating the high-frequency detail component obtained in the step 3

Absolute value of (2)

In particular, if

If it is not

$\left|D_{2^{-1}}f(x,y)\right|=-D_{2^{-1}}f(x,y);$

Step 5, counting the obtained in step 4

Is greater than a threshold value V_oNumber of pixel points of will

Is greater than a threshold value V_oAs boundary points of eyelids and eyelashes; v_oIn order to determine whether the pixel point is the threshold of the boundary point of the eyelids and eyelashes, the specific calculation formula is as follows:

where M and N are the width and height of the original normalized iris image,

step 6, comparing the values calculated in step 5

Value and threshold T_NComparing; if it is not

Then it is considered thatThe image is of the eyelid and eyelash obscuration if

The image is considered to be a processable normal iris image;

step 7, according to the method in step 3, passing the resolution 2 obtained in step 2^-2Lower horizontal direction wavelet coefficient

Vertical direction wavelet coefficientAnd diagonal direction wavelet coefficients

Reconstructing the original normalized Iris image at resolution 2^-2Details of

Step 8, calculating the result obtained in step 7

The variance of (a); the specific calculation formula is as follows:

wherein Var is resolution 2^-2The variance of the detail components, M and N, are the width and height of the original normalized iris image, respectively, and the calculation formula of M is: <math><mrow><mi>m</mi><mo>=</mo><munderover><mi>&Sigma;</mi><mrow><mi>x</mi><mo>=</mo><mn>1</mn></mrow><mi>M</mi></munderover><munderover><mi>&Sigma;</mi><mrow><mi>y</mi><mo>=</mo><mn>1</mn></mrow><mi>N</mi></munderover><msub><mi>D</mi><mrow><msup><mn>2</mn><mrow><mo>-</mo><mn>2</mn></mrow></msup><mi>f</mi><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow></mrow></msub><mo>/</mo><mi>M</mi><mo>&times;</mo><mi>N</mi><mo>;</mo></mrow></math>

step 9, the variance Var obtained in the step 8 and a threshold value T for judging the definition of the iris_vComparing, if Var is more than or equal to T_vIf so, the definition of the iris image meets the system requirement; if Var < T_vAnd the definition of the iris image does not meet the system requirement.

Through the steps, the normalized iris image extracted from the original image containing the iris can be analyzed to judge whether the image meets the requirements of the system.

It should be noted that:

1. the iris normalization operation in step 1 must be performed after iris localization.

2. Scale function for performing two-dimensional wavelet transform in step 2

Is a function of two one-dimensional scales

And

the product of (a). Horizontally sensitive wavelet function for two-dimensional wavelet transform

As a one-dimensional wavelet function psi_j，m(x) And a one-dimensional scale function

The product of (a); vertical sensitive wavelet function for two-dimensional wavelet transform

As a function of a one-dimensional scale

And a one-dimensional wavelet function psi_j，n(y) the product of (a); diagonally sensitive wavelet function for two-dimensional wavelet transform

For two one-dimensional wavelet functions psi_j，m(x) And psi_j，nThe product of (y).

3. Resolution 2 is selected in steps 3, 4, 5^-1The analysis of the eyelid and eyelash occlusion is performed for the detail components below because there are significant gray scale variations on the gray scale image at the edges of the eyelid and eyelash, and the points at these locations correspond to the larger amplitudes of the high frequency detail components according to the idea of multi-resolution analysis.

4. Threshold value V in step 5_oFor determining whether or not the points on the high-frequency detail components are edge points of the eyelids and eyelashes. Considering the magnitude of the high frequency detail component

Greater than V_oThe points (b) are the edge points of the eyelids and eyelashes because there is an obvious jump in the gray value at the edge positions of the eyelids and eyelashes, corresponding to the point with a larger amplitude on the high-frequency detail component; at the texture edge of the iris, the gray scale transformation is slower, corresponding to the point with smaller amplitude on the high frequency detail component. When the high frequency detail component amplitude

Above a certain value, the point is the edge point of the eyelids and eyelashes.

5. Threshold T mentioned in step 6_NIs an image for judging whether the image is an eyelid-free and eyelash occlusion image that can be processed. T is_NIs related to the subsequent matching algorithm, which we set as

Matching iris images of (1), T corresponding to the lowest error rate_N。

6. Step 7, 8, selecting the normalized iris image with the resolution of 2^-2Details of

The iris definition analysis is carried out because the iris texture changes slowly on the gray level image, and the change is reflected in the intermediate frequency detail component according to the idea of multi-resolution analysis

The clearer the iris image is, the larger the variance Var of the intermediate-frequency detail component is; the more blurred the iris image is, the smaller the variance Var of the mid-frequency detail component is.

7. Threshold value T in step 9_vIs used for judging whether the iris image is clear or not, T_vIs related to the subsequent matching algorithm.

The invention adopts a multi-resolution analysis method, firstly, the iris in the original gray level image is extracted and normalized; then, the normalized iris image is subjected to multi-resolution decomposition to respectively obtain the resolution 2^-1And resolution 2^-2The following detail components; finally according to resolution 2^-1Judging whether the image has eyelid and eyelash occlusion problems or not according to the number of larger amplitude points of the detail components, and judging whether the image has eyelid and eyelash occlusion problems or not according to the resolution 2^-2And judging whether the iris image is clear or not by the variance of the lower detail component. By adopting the method based on multi-resolution analysis, the quality of the iris image can be effectively evaluated, and the problem that the traditional algorithm is sensitive to illumination is avoided.

Firstly, compiling an iris image quality evaluation program by using C language; then, a CMOS or CCD camera device is adopted to automatically shoot the original image of the iris; then, inputting the shot iris original image as source data into an iris image quality evaluation program on a PC platform for processing; and judging whether the image meets the system requirements or not after iris positioning, normalization and image quality evaluation. 2400 gray iris images shot well and comprising different lighting conditions and different shooting postures of different people are used as source data, the result of program judgment is compared with the result of subjective judgment, the error probability is 1.2%, and the processing time of each image is less than 150 ms.

In conclusion, the method of the invention fully utilizes the texture information of the iris and combines a multi-resolution analysis method, thereby realizing the rapid and accurate judgment of the quality of the iris image.

Claims (2)

1. A method for judging quality of an iris image is characterized by comprising the following steps:

step 1, carrying out image acquisition on an iris in human eyes through a camera device, and obtaining a normalized iris image f (x, y) with the size of M multiplied by N from an original gray image containing the iris image; (x, y) represents the coordinates of the pixel points, and f (x, y) represents the gray value of the pixel points with the coordinates of (x, y);

step 2, performing 4-layer two-dimensional wavelet transform on the normalized iris image obtained in the step 1; specifically, the two-dimensional wavelet transform is formulated as：And

wherein,

is a resolution of 2^jThe scale factor of (a) is lower,

is a resolution of 2^jThe value range of the 4-layer wavelet transformation j is { -1, -2, -3, -4}, i { -H, V, D }, and details in the horizontal direction, the vertical direction and the diagonal direction are added;

in order to be a function of the scale,

selecting a wavelet as a wavelet function, wherein the selected wavelet is a DMEyer wavelet; scale function for performing two-dimensional wavelet transformIs a function of two one-dimensional scales

Andthe product of (a); horizontally sensitive wavelet function for two-dimensional wavelet transform

As a one-dimensional wavelet function psi_j，m(x) And a one-dimensional scale functionThe product of (a); vertical sensitive wavelet function for two-dimensional wavelet transform

As a function of a one-dimensional scale

And a one-dimensional wavelet function psi_j，n(y) the product of (a); diagonally sensitive wavelet function for two-dimensional wavelet transform

For two one-dimensional wavelet functions psi_j，m(x) And psi_j，n(y) the product of (a);

step 3, obtaining resolution 2 through step 2^-1Lower horizontal direction wavelet coefficientVertical direction wavelet coefficientAnd diagonal direction wavelet coefficients

Reconstructing the original normalized Iris image at resolution 2^-1Details ofI.e. high frequency detail components; specifically, the reconstruction formula is:

wherein

Representation resolution 2^jThe value of the detail component of the lower coordinate (x, y),

representation resolution 2^jThe wavelet coefficients of the lower one(s),

is a two-dimensional wavelet sensitive to the i direction; sigma is an accumulation operator;

step 4, calculating the high-frequency detail component obtained in the step 3

Absolute value of (2)

In particular, if

If it is not

$\left|D_{2^{-1}}f(x,y)\right|=-D_{2^{-1}}f(x,y);$

Step 5, counting the obtained in step 4Is greater than a threshold value V_oNumber of pixel points of will

Is greater than a threshold value V_oAs boundary points of eyelids and eyelashes; v_oIn order to determine whether the pixel point is the threshold of the boundary point of the eyelids and eyelashes, the specific calculation formula is as follows:

where M and N are the width and height of the original normalized iris image,step 6, comparing the values calculated in step 5

Value and threshold T_NComparing; if it is not

The image is considered as an image of the eyelid and eyelash occlusion if

The image is considered to be a processable normal iris image;

step 7, according to the method in step 3, passing the resolution 2 obtained in step 2^-2Lower horizontal direction wavelet coefficientVertical direction wavelet coefficient

And diagonal direction wavelet coefficients

Reconstructing the original normalized Iris image at resolution 2^-2Details of

Step 8, calculating the result obtained in step 7

The variance of (a); the specific calculation formula is as follows:

wherein Var is resolution 2^-2The variance of the detail components, M and N, are the width and height of the original normalized iris image, respectively, and the calculation formula of M is: <math><mrow><mi>m</mi><mo>=</mo><munderover><mi>&Sigma;</mi><mrow><mi>x</mi><mo>=</mo><mn>1</mn></mrow><mi>M</mi></munderover><munderover><mi>&Sigma;</mi><mrow><mi>y</mi><mo>=</mo><mn>1</mn></mrow><mi>N</mi></munderover><msub><mi>D</mi><mrow><msup><mn>2</mn><mrow><mo>-</mo><mn>2</mn></mrow></msup><mi>f</mi><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow></mrow></msub><mo>/</mo><mi>M</mi><mo>&times;</mo><mi>N</mi><mo>;</mo></mrow></math>

step 9, the variance Var obtained in the step 8 and a threshold value T for judging the definition of the iris_vComparing, if Var is more than or equal to T_vIf so, the definition of the iris image meets the system requirement; if Var < T_vAnd the definition of the iris image does not meet the system requirement.

2. The method as claimed in claim 1, wherein the threshold T is set in step 6_NIs an image for judging whether the image is a processable maskless eyelid and eyelash occlusion image, T_NIs associated with a subsequent matching algorithm, which is set to

Matching iris images of (1), T corresponding to the lowest error rate_N。

Images