KR20060111812A - Iris cognition method using comparison area embossed peculiar characteristic of iris pattern - Google Patents

Abstract

An iris recognition method using a comparison area embossing unique features of an iris pattern is provided to reduce a calculation time according to reduction of an iris feature code and increase an iris recognition ratio by recognizing only the comparison area embossing the unique iris feature. A preprocessed iris image of a user is divided into multiple areas to emboss structural feature of iris wrinkles represented in the iris image(S300). The iris feature code is generated by extracting and coding the structural feature of the iris wrinkles from each divided area(S320). A disparity of each divided area is calculated by exchanging a structural difference of the iris feature of each divided area into a distance value(S330). An area point of each divided area is output by calculating a disparity average of each divided area(S340). The area having a relatively high priority is selected as the comparison area by setting the priority to each divided area in area point order(S350). The selected comparison area and the iris feature code for the comparison area are stored to an iris feature database of the user(S500).

Description

Iris Recognition Method Using Comparative Areas with Unique Characteristics of Iris Patterns {IRIS COGNITION METHOD USING COMPARISON AREA EMBOSSED PECULIAR CHARACTERISTIC OF IRIS PATTERN}

1 is a flow chart showing a conventional iris recognition method.

Figure 2 is a flow chart schematically showing the concept of the iris recognition method according to the present invention.

Figure 3 is a flow chart showing in detail the iris recognition method according to the present invention.

4 is a flowchart illustrating a process of calculating region differences according to divisions according to the present invention.

Figure 5 is a model showing the concept of the area difference diagram and the calculation process of the area score according to the present invention.

6 is a flowchart illustrating a process of selecting a comparison area according to the present invention.

7 is a system configuration of the iris recognition method according to the present invention.

*** Explanation of the reference numerals for the main parts of the drawings ***

10: iris image 20: segmented area

30: area difference diagram 100: eye image acquisition module

200: preprocessing module 300: iris image segmentation module

400: feature code extraction module (registration process) 500: domain difference calculation module

600: area score calculation module 700: comparison area selection module

800: user's iris feature code registration module

900: feature code extraction module (identification process) 1000: feature code matching module

The present invention relates to an iris recognition method for identifying an identity using an iris in which a pattern is different for each person as one of biometric methods for identifying an individual by using a biological characteristic of a person. More specifically, in contrast to the prior art, in which the entire area of the iris is used as a recognition target area for recognizing the iris to identify an individual, the unique characteristic of the iris, in particular, according to the present invention, a comparative area in which the structural characteristics of the iris wrinkles are more highlighted The present invention relates to an iris recognition method which extracts and uses this as an area of an iris recognition target.

With the development of the information industry, as most important information is processed by computers, the demand for accurate personal authentication is growing more than ever, as the exchange of information through the Internet is active and the size of industries related to e-commerce such as online finance is growing. .

In line with this, biometrics that identify individuals by using the biological and behavioral characteristics of human beings are also rapidly spreading along with the drop in prices and rapid informatization according to the advancement of their own technology and the development of peripheral technology.

Biometrics is a technology that identifies an individual using the biometric and behavioral characteristics of a person. The fields are those that use biometric characteristics such as fingerprints, faces, irises, retinas, and vein recognition, as well as behaviors such as signatures and speech recognition. The field using characteristics is representative.

Among them, human iris has more than 1078 unique patterns and has abundant characteristic variables that can be distinguished from fingerprints having about 1010 patterns. And because of its uniqueness, since the independence of the individual is higher than the face or fingerprint, when applied to the identification system can have a high reliability and is widely used in the field of biometrics.

The iris is an area existing between the pupil and the white area of the eye. The iris is entangled with the pupil sphincter, the sympathetic muscle, the collagen fiber, the autonomic circumflex cavity, and forms a ring-shaped iris pattern with coloring. . The long band-like mesh (comb ligament) of this iris, corrugated red fibers, eyelash-like projections, sinusoidal vessels, ring-shaped circles, corona-shaped ligaments surrounding the pupil, The color, spots, etc. of the iris have different biological characteristics, and the shape is generated within about 18 months after the birth of the person, protected by the eyebrows, eyelids and retina, and the shape does not change easily during life. It is evaluated to have high consistency. Similarly, if you look closely at the iris of the eye, you can see that the patterns and shapes are different from person to person. And the iris is known to be identical for identical twins that have the same genetic structure as well as the left and right sides of the same person.

The iris recognition, which converts the wrinkles of the iris into a frequency and confirms its identity, has 250 discrimination factors, far ahead of 30 fingerprints. In addition, the iris of a living person has a slight tremor, which is clearly distinguished from that of the dead, and because of its round shape, misuse is practically impossible with flat images such as photographs and graphics.

In addition, the iris recognition using the iris pattern is capable of acquiring data at a certain distance or more from the eye without requiring physical contact of a finger with a sensor, such as fingerprint recognition, thereby allowing the iris recognition subjects to reject the use. This is less. In addition, after a certain period of time, the characteristics of the iris do not change for a lifetime. There is an advantage that there is no problem due to a change in biological characteristics such as a problem of feature extraction due to a deformation of the hair, a change in the length of the hair, or a facial expression.

Therefore, the iris recognition system using the above-described iris as a user authentication means has recently been attracting attention because it has a lower false recognition rate than any other system and is used where high security is required.

The fact that such an iris could be useful for personal identification was first proposed by Frank Burch in the 1930s. Later, American ophthalmologists Leonard Flom and Aran Safir derived research showing that iris patterns are unique to each person.In 1994, Professor John G. Daugman of the University of Cambridge, England, transformed Gabor Wavelet capable of encoding iris patterns into 256 bytes. A video signal processing algorithm based on (Gabor Wavelet Transform) is proposed, and all commercially available products are based on this algorithm.

On the other hand, the conventional iris recognition method for identifying an individual using the iris can be divided into the registration process of the iris recognition and the identification process of the iris recognition.

The registration process of iris recognition is to register the characteristic information of the iris image of the user who is the subject of personal identification. First, the user uses an image input device consisting of a camera for image acquisition, illumination for obtaining an accurate iris pattern, and an image sensor. Acquire an eye image (S10). In addition, an accurate iris region is extracted through a preprocessing process for detecting the inner boundary of the iris corresponding to the boundary between the pupil and the iris, and the outer boundary of the iris corresponding to the boundary between the sclera and the iris. ). Next, the feature of the pattern appearing in the extracted iris region is extracted and coded based on the Gabor Wavelet Transform, which is already commercially available (S30). Next, the extracted iris feature code is registered in the iris feature DB of the user (S40).

Thereafter, the identification process of the iris recognition extracts the iris region through pre-processing from the eye image acquired through the image input device similarly to the registration process of the iris recognition (S10 and S20). Next, since the feature of the pattern is extracted and coded for the extracted iris region, an iris feature code is generated (S30). Finally, the identity of the coded iris feature code and the iris feature code registered in the user's iris feature DB is measured to confirm whether or not the identity (S50).

In the iris recognition method through the above process, the main researches to date are the preprocessing process, the feature extraction process of the iris pattern, and the iris recognition means.

In the feature extraction process, the dimension of the feature vector is reduced to reduce the burden on the classifier of the iris pattern and to extract high-quality features to further improve the performance of the classifier in iris recognition rate and processing time. have. In this case, the iris pattern is characterized by physical location consisting of pixel position information that can be used in the pixel comparison method, size and size of pixel, etc., relative position of iris wrinkles, and relative position of structure composed of wrinkles and wrinkles. In addition, it can be divided into logical features including relationship information of image values existing in the iris image and relative position information of iris image elements.

In addition, as a study on the iris recognition means, a new structural recognizer based on k-NN, MLP, and tree comparison is used, and the image to be processed is classified into rectangular areas, and then the density of data pixels in each divided area is calculated. There is also a study to classify images using k-NN and MLP recognizer.

However, the previous studies listed above not only performed pixel-based computation for feature extraction, but also used the entire iris image for comparison. Therefore, as the iris feature is extracted and coded for the entire iris image, the calculation time for the large size of the iris feature code is considerable, which requires a large amount of processing time for the iris recognition.

However, until now, basically, the feature is extracted from the entire iris image and recognition is performed. As a method for extracting better features, there is no research for extracting a comparison area.

Therefore, the present invention was devised in view of the above technical problems, and thus, when the iris recognition is performed, the target area is performed only for the comparison area where the iris inherent characteristics are highlighted. The purpose of the present invention is to provide an iris recognition method having an improved iris recognition rate because the iris recognition is performed only on a region including high quality features.

In order to achieve the technical problem as described above, the present invention, in the iris recognition method for extracting the iris image by performing preprocessing on the eye image of the user obtained through the image input device and identifying the identity of the user from the ( a) dividing the iris image of the user who has undergone the pre-processing into a plurality of areas so that the structural features of the iris wrinkles shown therein are highlighted; (b) dividing the structural features of the iris wrinkles by the regions divided in the step (a) Extracting and encoding the iris feature code to generate an iris feature code. Computation of the degree, (d) the divided region in which the average value of the region difference is calculated for each divided region for all the iris images calculated by the step (d) Calculating the area scores for each area, (e) setting priorities for the divided areas in order of the area scores for each area calculated in the step (d), and comparing the divided areas with the higher priority as the comparison area. Selecting (f) registering the selected comparison area and the iris feature code thereof in the iris feature DB of the user, and (g) registering the iris feature information of the user through the steps (a) to (f) Subsequently, extracting only the iris feature of the comparison area selected in step (e) for the iris image of the user who wishes to confirm the identity request after preprocessing, generating an iris feature code, and (h) generating the iris feature code in step (g). Since the similarity is measured by performing matching between the iris feature code and the iris feature code registered in the user's iris feature DB in step (f), identifying the identity of the user.

Step (c) is to generate all possible pairs for all iris images having undergone steps (c1), (a) and (b), and (c2) for each segment of each iris image of the pair. Obtaining a feature vector for a code and calculating a distance value between the feature vectors, (c3) generating a weight graph by applying the converted distance value to a dynamic program matching method, (c4) in the generated weight graph The method may include finding a shortest path and calculating an area difference obtained by summing all weight values corresponding to the path.

And the area score in the step (d) is a value that reflects this by expressing the area size of the noise and segmentation area other than the iris wrinkle information included in the iris feature information due to external factors such as eyelashes, eyelids, lighting, etc. do.

In the step (e), (e1) assigning a rank based on the area score of each divided area, (e2) recognizing the iris wrinkle pattern for the divided area having a high priority, and (e3) having a high priority Repeating step (e2) for each segment in order and determining whether the recognition rate of the iris wrinkle pattern reaches a threshold where the recognition rate does not increase any more, and (e4) step (e3), the recognition rate reaches the threshold. And determining a plurality of comparison areas in order of high priority for all the partitions recognized up to that point.

Hereinafter, an embodiment of an iris recognition method using a comparison area in which the inherent characteristics of the iris pattern of the present invention is highlighted will be described in detail with reference to the accompanying drawings.

2 is a flow chart schematically showing the concept of the iris recognition method according to the present invention. As shown in the drawing, according to the present invention, in the registration process of the iris recognition, the region for extracting the feature information on the iris image of the user is specified, and in the identification process of the iris recognition, the feature information of the iris is identified only for the specific region. The main purpose is to extract and determine the identity of the user. In particular, the present invention extracts and uses the structural features of the wrinkles of the iris as the characteristic information of the iris.

Referring to the series of processes of the iris recognition method according to the invention as follows.

First of all, the user can use a video input device such as a CCD (Change Coupled Device) camera, a video camera, illumination to obtain a clear iris pattern, and a capture board or frame grabber to convert an analog image into a digital image. Acquire an eye image (S100).

Next, a preprocessing process for iris recognition is performed on the obtained eye image. The preprocessing process is to extract only the iris region from the acquired eye image. The inner boundary of the iris corresponding to the boundary between the pupil and the iris and the outer boundary of the iris corresponding to the boundary of the sclera and iris, which are white Since it detects the correct iris area (S200).

Next, in the process of registering the user's iris feature information in the iris recognition system, an area in which the structural features of the iris wrinkles are highlighted for the extracted iris image is selected as a comparison area (S300).

Next, feature information of the iris shown in the comparison area of the iris image is extracted (S400). At this time, the feature information of the iris is coded by extracting the feature of the iris wrinkle pattern appearing on the comparison area.

Next, the extracted iris feature information is registered in the iris feature DB of the user (S500).

After the registration process for the iris recognition as described above, the identification process of the iris recognition is performed.

As in the registration process, first, as in the registration process, first, an eye image of the user requesting identification is obtained (S100), and only the iris region is extracted through a preprocessing process for the acquired eye image (S200).

Next, only the characteristic information of the iris shown in the region selected as the comparison region is extracted with respect to the extracted iris image (S600).

Finally, matching of the extracted iris feature information with the iris feature information registered in the iris feature DB of the user is performed to confirm the identity of the user (S700).

As described above, when the feature information of the iris is extracted, it is common to extract and code the feature value for the entire image of the iris image. However, according to the present invention, a comparison region from which the feature information of the iris is extracted is selected, Extracting only the characteristic information of the iris wrinkles shown and performing matching of the iris features only with respect to the comparison area for later identification is the main part of the present invention.

Figure 3 is a flow chart showing in detail the iris recognition method according to the present invention. As shown in the drawings, a series of processes of the iris recognition method according to the present invention are as follows.

First, only the iris region is extracted by performing preprocessing on the eye image acquired by the image input apparatus (S100 and S200).

And according to the registration process of the iris recognition, the pre-processing (S200) is divided into a plurality of areas so as to highlight the structural features of the iris wrinkles for all the iris image to be registered in the user's iris feature DB (S310).

Next, the structural features of the iris wrinkles are extracted and coded for each of the divided regions, thereby generating an iris feature code (S320).

Subsequently, a region difference degree indicating a difference in structural characteristics of the iris wrinkles extracted in the step S320 is calculated for each segment (S330). Here, the region difference degree will be described in more detail with reference to FIG. 4.

Next, the area score for each divided area is calculated by reflecting the area difference degree calculated in the step S330 (S340). In this case, the area score is a value obtained by adding an area difference degree value between all paired iris images for each divided area and calculating an average thereof. See Equation 1 below.

Savg (i): Area score for the i th partition

_N C ₂ : Total number of pairs for a paired iris image (n: Number of all iris images)

d _Nij : All pairwise region difference values for the i th partition

∑d _Nij : Sum of all pairwise region difference values for the i-th partition

However, when the comparison area is selected later, if noise exists in the divided region or the iris wrinkles are incorrectly extracted, it is a major cause of the false recognition rate. Here, the noise is other than iris wrinkle characteristic information caused by external factors such as eyelashes, eyelids, and lights when acquiring an image of an eye, and the likeness of wrinkles, the disappearance of wrinkles, one wrinkle is divided into two or more wrinkles, and a part of wrinkles Such as extinction.

In addition, the size of the area divided in the step (S310) varies depending on the characteristics of the iris wrinkles. That is, as the area is large and small, the feature extraction of the iris wrinkles may or may not be easier, which is a factor influencing iris recognition.

Therefore, the noise information and the size information of the divided region are calculated as a penalty function and reflected in the calculation of the region score for each divided region. The penalty function is shown in Equation 2 below.

β ₁ ,: Proportional constant representing the noise information reflection ratio in the penalty function

β ₂ : Proportional constant representing the size reflecting ratio of the partition to the penalty function

a _i : Area size value of the i th partition

a _m : the maximum area value of the segment (ie the total area of the iris image including the segment)

r: value of noise information included in partition

That is, the penalty function value is an area size value of a divided area that is an object of area score calculation.

The larger (a _i ) (ie, the closer to the maximum area value of the partition) As the value of close to 0, the penalty function becomes smaller, which is inversely proportional to the area size of the partition. As the noise information value r increases, the penalty value increases, which is proportional to the noise information value. When the penalty value is reflected in the area score, Equation 3 is obtained.

S _i = s _{avg (i)} -p _i

S _i : Area score value for the i-th partition that reflects the penalty function value

s _{avg (i)} : Area score value for the i th partition

p _i : Penalty function of noise and area size information for the i-th partition

That is, the calculated noise information and area size information for the i-th partition to the penalty function, in an area score value calculated by the average value _{(s avg (i))} of the area dissimilarity of the i-th partition value (p _i ) Is the area score for the i-th partition.

Next, a segmented region having a relatively high region score calculated in step S340 is selected as a comparison region (S350). This will be described in more detail with reference to FIG. 6.

As described above, the process of selecting the comparison area (S300) where the structural features of the iris image are highlighted by performing the steps S310, S320, S330, S340, and S350.

Then, the feature code of the iris wrinkles for the comparison area selected in the step (S350) is registered in the user's iris feature DB (S400).

On the other hand, after the registration process for the iris recognition is performed, the iris recognition process for the user who wants to verify the identity is the comparison area selected in the step (S350) for the iris image of the user who went through the pre-processing (S100, 200) By extracting only the features of the iris wrinkles shown in the coded it generates an iris feature code (S500).

Then, by matching the iris feature code extracted in the step (S500) and the iris feature code registered in the user's iris feature DB in the step (S400) to measure the similarity is to identify the user's identity (S600).

FIG. 4 is a flowchart illustrating a process of calculating region differences for each divided region according to an exemplary embodiment of the present invention, and step S330 of FIG. 3 will be described in more detail.

As shown in the figure, the process of calculating the area difference degree according to the present invention (S330) is first arranged for all possible pairs using the combination formula described in Equation 4 for all the iris image pre-processed (S331) ).

_N C ₂

That is, Equation 1 selects two iris images for n iris images.

Subsequently, the structural difference of the characteristic of the iris wrinkles displayed in the divided region between the paired iris images in step S331 is converted into a distance value using Equation 5 below (S332).

d (i, j) = | a _li- b _lj | + α (| a _s1- b _s2 | + | a _e1- b _e2 |)

a: starting point of iris wrinkle

b: end point of iris wrinkle

α: proportional constant that indicates how much the distance between the start and end points of the iris wrinkles is reflected in the formula

i, j: index of the iris feature vector

That is, the structural characteristics of the iris wrinkles on the divided regions between the paired iris images are converted into distance values between the start point and the end point of the iris wrinkles. In more detail, for example, for the feature vectors for each iris wrinkle feature of the paired iris image, there are two index types, m and n, respectively, 1 ≦ i ≦ m, 1≤j≤n is satisfied. And d (i, j) means the distance between the line segments between the feature vectors of the paired iris image. Therefore, the feature vectors of the paired iris images are ((al1, as1, ae1), (al2, as2, ae2), ..., (alm, asm, aem)) and ((bl1, bs1, be1), ( For bl2, bs2, be2),…, (bln, bsn, ben)), the distance between the i-th segment and the j-th segment is d (i, j). The calculation is then made taking into account the length.

Next, the distance value converted in the step S332 is reflected in the dynamic program matching method to generate a weight graph (S333). The dynamic program matching method is the most efficient method of matching patterns regardless of the number of elements of each pattern or two elements other than pattern feature elements such as noise. The formula is shown in Equation 6 below.

i, j: index of the iris feature vector

C: Dynamic Program Matching Constant

Next, the shortest path among the weight values in the weight graph generated in step S333 is detected, and a value obtained by adding up all the weight values corresponding to the shortest path is defined as an area difference (S334). That is, the formula for calculating the region difference degree is expressed by Equation 7 below.

w (i): i-th weight value on the weight graph

FIG. 5 is a diagram illustrating a concept of an area difference diagram and an area score calculation process according to the present invention.

As shown in the figure, if there are three (A, B, C) iris images 10, possible pairs also include (A, B), (B, C), and (C, A). In addition, all the iris image 10 is divided into a number of areas where the characteristics of the iris wrinkles are well highlighted, for example, it is assumed to be divided into n.

Then, n region difference values may be obtained for the three pairs (A, B), (B, C), and (C, A) of the iris image 10. Here, the method for calculating the region difference value has already been described with reference to FIG. 4. In this case, the n area difference degree values obtained for each pair are referred to as the area difference diagram 30. The region difference values of the divided regions corresponding to each other between the paired iris images on the region difference diagram 30 for each of the three pairs are called d1, d2, and d3, respectively. It means the region difference value between the iris image 10 forming the first pair A, B, the second pair B, C, and the third pair C, A with respect to the divided region 20. An area difference degree vector may be obtained using the area difference degree values, and a total of n area difference degree vectors d1, d2, and d3 may be obtained for each partition.

이 된다.Then, after adding all the region difference values for each of the divided regions with respect to the n region difference degree vectors d1, d2, and d3, the average value is calculated as the region score for each region. That is, the area score for the divided area 20 shown in FIG.

Becomes

6 is a flowchart illustrating a process of selecting a comparison region according to the present invention. As shown in the drawing, in the selection process of the comparison area (S350), first, the priority is given to the divided area having the high area score based on the area score calculated for each divided area (S351).

Next, the iris wrinkle pattern is recognized for the high priority division area set in step S351 (S352). Herein, the recognition of the iris wrinkle pattern is performed by a commonly used iris recognition method such as a Dobechies wavelet transform and a Harb wavelet transform.

Then, it is determined whether the recognition rate for the iris wrinkle pattern in the step (S352) reaches a threshold which no longer increases, and the above step (S352) is repeated until the threshold is reached (S353). The threshold here refers to whether the recognition rate of the iris feature information is close to 100.

Next, since the recognition rate no longer increases through the step S353, when it is determined that the threshold is reached, a plurality of comparison areas are adopted in order of high priority for all the divided areas recognized up to that time (S354).

7 is a system configuration of the iris recognition method according to the present invention. As shown in the drawings, the present invention, the eye image acquisition module 100, pre-processing module 200, iris image segmentation module 300, feature code extraction module 400 in the registration process of the iris recognition, area difference diagram Calculation module 500, area score calculation module 600, comparison area selection module 700, user's iris feature DB registration module 800, feature code extraction module 900 in the identification process of the iris recognition, features Code matching module 1000 is configured.

The eye image acquisition module 100 and the preprocessing module 200 obtain an eye image of a user by an image input device and extract only an iris region therefrom.

The iris image segmentation module 300 divides the iris image extracted by the eye image acquisition module 100 and the preprocessing module 200 into regions in which structural features of the iris wrinkles are highlighted.

The feature code extraction module 400 in the registration process of the iris recognition extracts and codes the feature of the iris wrinkles displayed on the area segmented by the iris image segmentation module 300.

The area difference degree calculation module 500 calculates an area difference degree indicating a structural difference of the iris characteristics for each segment based on the iris feature information coded by the feature code extraction module 400 during the registration process of the iris recognition. Calculate

The area score calculation module 600 calculates an average value for each of the divided areas with respect to the area difference degree calculated by the area difference degree calculation module 500, and uses this as the area score for each divided area.

The comparison area selection module 700 ranks the area scores calculated by the area score calculation module 600 in the order of the highest scores for each area and selects a relatively high ranking area as the comparison area. .

The user's iris feature DB registration module 800 registers the iris feature information for the comparison area selected by the comparison area selection module 700 in the user's iris feature DB.

The feature code extraction module 900 in the identification process of the iris recognition is performed by the comparison region selection module 700 for the iris image of the identification subject extracted by the eye image acquisition module 100 and the preprocessing module 200. Only the features of the iris wrinkles in the selected comparison area are extracted and encoded.

The feature code matching module 1000 matches the iris feature information coded by the feature code extraction module 900 and the iris feature information registered by the user's iris feature DB registration module 800 in the identification process of iris recognition. To identify the identity of the person being identified.

As described above, the present invention selects a comparison region in which iris wrinkles are structurally different and the characteristics of the iris are highlighted without using the entire image of the iris image in iris recognition. By comparing, the calculation time due to the size reduction of the iris feature code to be extracted is shortened, and ultimately the time required for iris recognition is also reduced. In general, the iris wrinkles are complex in the iris pattern, and similar iris images are similarly determined only in regions where the characteristics of the patterns are significantly different, thereby improving the iris recognition rate.

In addition, when the eyelashes are covering the pupils or the eyelids cover a large part of the iris during the acquisition of the iris image, the enlargement or reduction of the pupil due to excessively dark or bright lighting, or the reflection of the light generated on the pupils, such as when the image is acquired The noise information other than the iris image is reflected in the selection of the comparison area to avoid the reflected area, thereby providing an iris recognition method that is strong in noise and has an accurate recognition rate.

Although the present invention described above has been described in detail only with respect to the specific examples described, it will be apparent to those skilled in the art that various modifications and changes are possible within the technical scope of the present invention, and such modifications and modifications belong to the appended claims. .

Claims (4)

In the iris recognition method for extracting the iris image by performing pre-processing on the user's eye image obtained through the image input device and identifying the user's identity therefrom, (a) dividing the iris image of the user who has undergone the pre-processing into a plurality of regions so that the structural features of the iris wrinkles shown therein are highlighted; (b) generating an iris feature code by extracting and coding the structural features of the iris wrinkles for each of the divided regions in step (a); (c) calculating a region difference degree for each divided region by converting the structural difference of the iris characteristics into distance values between the iris images which have undergone the steps (a) and (b); (d) calculating area scores for each divided area in which the average value of the area difference is calculated for each divided area for all the iris images for which the area difference degree is calculated in the step (d); (e) setting priorities for the divided regions in order of the area scores of the divided regions calculated in the step (d) and selecting the divided regions having the higher priority as the comparison region; (f) registering the selected comparison area and an iris feature code thereof with the iris feature DB of the user; (g) After registering the iris feature information of the user through the steps (a) to (f), only the iris feature of the comparison area selected in the step (e) for the iris image of the user who desires the identity verification request that has been preprocessed. Extracting to generate an iris feature code; And (h) identifying the identity of the user because the similarity is measured by performing matching between the iris feature code generated in step (g) and the iris feature code registered in the user's iris feature DB in step (f). An iris recognition method using a comparison area in which the unique characteristics of the iris pattern are highlighted. The method of claim 1, wherein step (c) (c1) generating all possible pairs for all iris images which have passed through steps (a) and (b); (c2) obtaining a feature vector for the feature code of each paired iris image for each divided region and calculating a distance value between the feature vectors; (c3) generating a weight graph by applying the converted distance value to a dynamic program matching method; (c4) finding a shortest path in the generated weight graph and calculating an area difference degree obtained by summing all weight values corresponding to the path; and obtaining a comparison area in which the unique characteristics of the iris pattern are highlighted. Iris recognition method used. The method of claim 1, wherein the area score in the step (d) is Comparative area that highlights the unique characteristics of the iris pattern, characterized by a penalty function that expresses the noise and segment area size other than the iris wrinkle information included in the iris feature information due to external factors such as eyelashes, eyelids, and lights Iris recognition method using a. The method of claim 1, wherein step (e) (e1) assigning a rank based on the area score of each divided area; (e2) recognizing an iris wrinkle pattern for the partition having a high priority; (e3) repeating step (e2) for each divided region in the order of high priority and determining whether the recognition rate of the iris wrinkle pattern reaches a threshold at which the recognition rate no longer increases; And (e4) if it is determined that the recognition rate reaches a threshold through step (e3), selecting a plurality of comparison areas in order of high priority for all the divided areas recognized up to that time; An iris recognition method using a comparison area in which the inherent characteristics of a pattern are highlighted.

Images