KR102329128B1 - An adaptive quantization method for iris image encoding - Google Patents

Abstract

A user recognition method using an iris, comprising: generating a first mask that blocks a non-iris object region of an iris image; generating a transformed iris image in which a non-iris object region is blocked according to the first mask; adaptively transforming the first mask according to the properties of , generating a second mask that additionally blocks a non-constant region in which the quantization result of the transformed iris image is not constant; quantizing pixels included in the iris image; Obtaining a code, applying a second mask to the iris code to obtain a transformed iris code in which a portion corresponding to a non-iris object region and a non-regular region is blocked, and transformation with a reference iris code stored in advance by a user A user recognition method using an iris, comprising the step of recognizing a user by matching an iris code.

Description

AN ADAPTIVE QUANTIZATION METHOD FOR IRIS IMAGE ENCODING

The present invention relates to iris recognition, and more particularly, to an adaptive quantization method for iris image encoding.

The electronic device may store information related to personal information such as personal location information, memo, and financial transaction, as well as information such as contact information, call logs, and messages. In order to protect the above personal information, the electronic device may be provided with various security functions. In particular, a method of maintaining the security of an electronic device by using the user's biometric information is widespread. Methods of maintaining the security of an electronic device using such biometric information include fingerprint recognition, face recognition, iris recognition, and the like.

In particular, iris recognition is an authentication technology for security purposes that uses characteristics of different iris between individuals. In addition, iris recognition may be performed using a camera without direct body contact.

However, the use of iris recognition technology in mobile devices may, for example, be caused by severely changing environmental conditions (e.g. indoor/outdoor, sunny/overcast weather, whether glasses/contact lenses are worn), when the device is in real-time mode. It is associated with numerous issues and difficulties, such as device performance limitations (CPU, RAM, camera resolution, etc.) that prevent iris recognition from being performed, and difficulty in user interaction (interaction is sometimes not convenient for users).

A conventional workflow of iris recognition involves quantization of iris images for iris code matching. However, for example, a change in environmental conditions or a non-constant bit appearing in quantization increases a false reject rate (FRR) and deteriorates iris recognition accuracy and robustness. Accordingly, it is an object of the present invention to search for and remove these non-constant bits using an adaptive threshold during or before iris code matching or database registration in order to improve the accuracy and robustness of iris code matching.

SUMMARY OF THE INVENTION The present invention solves the problems and disadvantages detailed below and provides at least the advantages described below.

Certain portions of the iris image may provide non-constant visual information. Accordingly, the performance of iris recognition may be deteriorated due to the portion providing non-constant visual information. Therefore, in order to improve the performance of iris recognition, it is necessary to block a portion providing inconsistent visual information in the iris recognition process.

A user recognition method using an iris, comprising: generating a first mask that blocks a non-iris object area of an iris image;

generating a transformed iris image in which the non-iris object region is blocked according to the first mask; adaptively transforming the first mask according to a property of the transformed iris image so that the quantization result of the transformed iris image is constant generating a second mask that additionally blocks an inconsistent inconsistent area; obtaining an iris code by quantizing elements included in the iris image; obtaining a transformed iris code in which the non-iris object region and a portion corresponding to the non-regular region are blocked by applying There is provided a user recognition method using an iris comprising the step of:

In a user recognition apparatus using an iris, a first mask is generated for blocking a non-iris object area of an iris image, and the non-iris object area is blocked according to the first mask. a second mask that additionally blocks an inconsistent area in which a quantization result of the transformed iris image is inconsistent by generating A mask generator that generates a mask, obtains an iris code by quantizing elements included in the iris image, and applies the second mask to the iris code to correspond to the non-iris object region and the non-constant region A user recognition device using an iris is provided, comprising: an iris code acquisition unit for acquiring a partially blocked transformed iris code; do.

A computer-readable recording medium in which a program for executing a user recognition method using an iris according to various embodiments provided herein is recorded.

1 is a flowchart illustrating an iris recognition process according to an embodiment of the present invention.
2 shows a method of extracting an iris image from an eye image.
3 and 4 show examples of normalization of iris images.
5 shows an example of a first mask and an iris image to which the first mask is applied.
6 illustrates a transformed iris image in which a non-iris object region is blocked by applying a first mask.
Fig. 7 discloses a transformed iris image that has been binarized in order to emphasize the characteristics of the iris.
Fig. 8 shows the generation of an iris feature vector according to a Gabor filter and binarization.
9 illustrates a description of a non-constant region in relation to an iris feature vector.
10 shows a specific example of determining a threshold value according to a blocking quota.
11 illustrates an example of generating a second mask by transforming a first mask according to a property of a transformed iris image.
12 shows a specific example of obtaining a transformed iris code by applying a second mask to the iris code.
13 illustrates a user recognition apparatus according to an embodiment of the present invention.
14 illustrates a user device including a user recognition device according to an embodiment of the present invention.
In the following description, the same reference numerals are used for the same elements, and overlapping descriptions are omitted unless described in other drawings.

The following description provided with reference to the accompanying drawings is provided to provide a comprehensive understanding of the various embodiments of the present application defined by the claims and their equivalents. Various specific details included to aid understanding are to be regarded as illustrative only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the various embodiments described herein can be made without departing from the scope of the disclosure. In addition, descriptions of well-known functions and configurations may be omitted for clarity and conciseness.

The terms and words used in this specification and claims are not limited to the bibliographical meaning, but are merely used by the inventors to enable a clear and consistent understanding of the various embodiments disclosed herein. Accordingly, it will be apparent to those of ordinary skill in the art that the following description of various embodiments disclosed herein is provided for purposes of illustration only.

Terms and words in the singular form are to be construed to include the plural unless the context clearly dictates otherwise. For example, unless otherwise specified, “eye region” is construed as indicating one or more of a plurality of images for “eye region”.

Although terms indicating the order of things, such as 'first' and 'second', may be used for the components of the present invention, the components are not limitedly interpreted by the terms as described above. Such terms are used only to distinguish one component from another.

As used herein, the term "comprising" indicates the presence of a specified feature, operation, element, and does not exclude the presence of one or more other unspecified feature, operation, element, or group thereof.

In various embodiments of the present disclosure, a “module” or “unit” may perform at least one function or operation, and may be implemented in hardware, software, or a combination thereof. “A plurality of modules” or “a plurality of units” may be implemented with at least one processor (not shown) through integration with at least one module other than a “module” or “unit” that needs to be implemented with specific hardware. have.

Hereinafter, various embodiments of the present invention will be described in more detail with reference to the accompanying drawings.

The electronic device according to various embodiments of the present disclosure may include a smart phone, a tablet PC (Personal Computer), a mobile phone, a video phone, an e-book reader, a desktop PC, a notebook PC PDA, a PMP, an MP3 player, a mobile medical device, a camera, or wearable devices (eg, head mounted displays (HMDs), eg, electronic glasses, electronic clothing, electronic bracelets, electronic necklaces, electronic accessories, electronic tattoos, or smart watches).

1 is a flowchart illustrating an iris recognition process according to an embodiment of the present invention. The iris recognition process according to an embodiment of the present invention may include the following steps S101 to S106.

In step S101, a first mask for blocking a non-iris object area of the iris image is generated. In order to obtain an iris image, the eye image may be extracted from a face image captured by an internal or external camera of the electronic device. The face image refers to an image of the user's entire face or at least a part of the face including the eye region. The eye image corresponds to an area of the eye in the face image. The electronic device for acquiring the face image includes, but is not limited to, a mobile terminal, a mobile phone, a table computer, a smart watch, and the like. The face image may be captured in a colored tone or a non-colored tone according to an embodiment.

The iris image is an image of the iris included in the eye and objects around the iris. The iris image is segmented from the eye image, and the term “segment” herein means highlighting or selecting a specific object from the image. The iris image may include objects such as pupil, eyelid, eyelash, sclera, and iris. An example of iris image segmentation is shown in FIG. 2 .

2 shows an eye image 210 and an iris image 220 . A method of segmenting the iris image 220 from the eye image 210 included in the face image 200 is illustrated on the left side of FIG. 2 . In addition, the iris image 220 highlighted in gray according to the result of the division is shown on the right side of FIG. 2 . Referring to the left side of FIG. 2 , the region of the iris image 220 is determined by excluding the circular region determined according to the size of the pupil from the circular region determined according to the size of the pupil. Thus, the region of the iris image 220 is ring-shaped.

The pattern of the iris is determined by analyzing the muscles of the iris radiating from the pupil. Accordingly, when a circular iris image is converted into a rectangular iris image, analysis of the iris pattern may be easier. The transformation of the iris image as described above is defined as normalization of the iris image. 3 and 4 illustrate normalization of the iris image.

According to the left side of FIG. 3 , the position of each pixel in the annular iris image 300 may be expressed as (r, θ) according to the polar coordinate system. In (r, θ), r means the distance of the pixel from the center, and θ means the direction of the pixel from the center. The annular iris image 300 can be normalized by corresponding r to the vertical axis and θ to the horizontal axis, as shown on the right side of FIG. 3 . For example, A and B displayed in the circular iris image 300 correspond to A and B displayed in the rectangular iris image 310 . The other pixels of the annular iris image 300 also correspond to the rectangular iris image 310 according to (r, θ). A rectangular normalized iris image 310 is thus created.

4 shows an example of iris image normalization. An annular iris image 400 is shown on the left side of FIG. 4 . In addition, an iris image 410 in which an annular iris image is normalized is shown on the right side of FIG. 4 . As in an example of the normalized iris image 410 of FIG. 4 , when an eyelid or eyelashes cover a portion of the iris, the performance of iris recognition may be deteriorated. Therefore, it is necessary to remove obstructions such as eyelids and eyelashes from the iris image. In this specification, a region corresponding to an obstruction such as an eyelid or eyelashes is defined as a non-iris object region. Conversely, the region corresponding to the iris is defined as the iris object region.

In order to remove the non-iris object region, a mask for inactivating the non-iris object region may be generated. In this specification, a mask for removing the non-iris object region is defined as a first mask. However, the name of the first mask is only arbitrarily defined for convenience of description, and may be defined as another name.

5 shows an example of a first mask 500 . The black portion of the first mask 500 corresponds to the non-iris object region. Accordingly, the portion of the iris image corresponding to the black portion is not used in iris matching. The white portion of the first mask 500 corresponds to the iris object region. Accordingly, the portion of the iris image corresponding to the white portion may be used to match the corresponding portion of the pre-stored reference iris image. In the present specification, the reference iris image refers to an image showing the user's iris pattern stored in advance in an internal or external database.

The first mask may be expressed in a bit form. Whether a pixel of the iris image is blocked may be expressed by at least one bit. When the 1 bit is 0, the pixel corresponding to the 1 bit is blocked from iris matching. Conversely, when the 1 bit is 1, the pixel corresponding to the 1 bit may be used in iris matching. If each pixel of the iris image is expressed as a complex number including a real part and an imaginary part, whether or not a pixel of the iris image is blocked may be expressed by 2 bits.

The first mask may be equally applied not only to the iris image but also to the reference image compared with the iris image. As another example, when there is a reference mask for removing the non-iris object region of the reference image, a merge mask in which the reference mask is merged with the first mask may be equally applied to the iris image and the reference image.

In step S102, a transformed iris image in which the non-iris object region is blocked according to the first mask is generated. 6 illustrates a transformed iris image 600 in which the non-iris object region is blocked by applying the first mask. The non-iris object region is blocked in the transformed iris image, so that only the iris image of the iris object region remains.

The iris pattern of the transformed iris image may be quantized for iris matching. Quantization is to convert a physical quantity having a continuous value into a physical quantity having a discontinuous value. The quantization step may include a binarization step of representing numerical data as bits having values of 0 and 1, and may be a part of the encoding process. As an example of binarization, when the pixel value of the transformed iris image is less than 0, the pixel value may be determined as 0, and when the pixel value of the transformed iris image is greater than 0, the pixel value may be determined as 1. Therefore, a pixel as a result of binarization can be expressed in the form of bits.

The iris pattern of the same user may be photographed slightly differently depending on the user's health and photographing environment. If there is a difference between the iris pattern of the reference iris image and the iris pattern of the iris image due to the influence of the shooting environment, etc., the user authentication device may recognize the iris image originating from the iris of the same user as an iris image originating from the iris of another person. . Therefore, by binarizing the transformed iris image, it is possible to remove minute differences in iris patterns generated from the photographing environment.

However, when the pixel value of the transformed iris image is close to 0, the binarization result of the pixel value may not be constant. For example, a pixel having a pixel value between -1 and 1 may be binarized as 0 or binarized as 1 according to the surrounding environment. In the present specification, a bit of a pixel whose binarization result is not constant is defined as an inconsistent bit. Therefore, for robustness of iris matching, it is necessary to additionally block pixels determined to include non-constant bits in the iris matching process.

7, a binarized transformed iris image 700 is disclosed. The gray portion of the binarized transformed iris image 700 represents a region binarized with a value of 1. And the white part of the binarized transformed iris image 700 represents a region binarized with a value of zero. Since pixels located at the boundary between the gray part and the white part have a high probability of having a value close to 0, the pixel value may be binarized to 0 or binarized to 1 depending on the surrounding environment. Therefore, after determining whether the binarization result is constant with respect to pixels located at the boundary between the gray portion and the white portion, it is necessary to additionally block the pixels determined to have non-constant binarization results from iris matching.

In step S103, by adaptively transforming the first mask according to the properties of the transformed iris image, a second mask is generated that additionally blocks an inconsistent area in which the result of quantization or binarization of the transformed iris image is not constant. . The non-constant region includes pixels whose quantization or binarization result is highly likely to change depending on the surrounding environment. The non-constant area is determined based on pixels of the transformed iris image, and the first mask is transformed to additionally block the non-constant area as well as the non-iris object area. A second mask that blocks both the non-iris object area and the non-uniform area is generated according to the conversion result.

Each pixel of the transformed iris image may have a real value. The real value of the pixel may be determined from the gray intensity corresponding to the pixel of the iris image.

Also, each pixel of the transformed iris image may be represented by a complex number. A pixel value expressed as a complex number may be determined from a gray intensity corresponding to a pixel of the iris image. According to an embodiment, the gray intensity of pixels of the transformed iris image may be expressed as a set of complex numbers indicating amplitude information and phase information of the iris pattern according to a Gabor filter. In iris recognition, amplitude information affected by a light source or a camera gain is removed, and only phase information indicating the sign of a complex number corresponding to pixels of the transformed iris image may be used. In addition to the Gabor filter, the phase information may be extracted by other techniques or filters including a filter based on a Fourier transform, a filter based on a wavelet transform, and the like.

8 shows a graph 800 of a method of determining an iris feature vector according to a Gabor filter and binarization. As described above, according to the Gabor filter, the gray intensity of a pixel can be expressed as a complex number. In iris recognition, amplitude information indicating the magnitude of the complex number is removed, and only phase information indicating the sign of the complex number is used. Accordingly, the iris feature vector may be generated by binarizing the complex value of the pixel obtained by the complex Gabor filter.

An equation for generating the iris feature vector described in the graph 800 is as follows.

if Re > 0, Im > 0 then {1, 1}

if Re > 0, Im < 0 then {1, 0}

if Re < 0, Im < 0 then {0, 0}

if Re < 0, Im > 0 then {0, 1}

In the above formula, Re denotes the real part of the complex value, and Im denotes the imaginary part of the complex value. For example, when the pixel value is 3.8+2.5j, since both the real and imaginary parts of the pixel value are greater than 0, the iris feature vector is determined as {1, 1}. As another example, when the pixel value is -4+1.2j, since the real part of the pixel value is less than 0 and the imaginary part is greater than 0, the iris feature vector is determined as {-1, 1}. An iris code, which will be described later, is determined as a set of iris feature vectors.

When pixels of the transformed iris image are expressed as a complex number, the non-constant region of the transformed iris image may include a first non-constant region for the real part of the pixels and a second non-constant region for the imaginary part of the pixels. The first non-constant region includes pixels in which the real part is close to zero and the result of binarization of the real part is not constant. In addition, the second non-constant region includes pixels whose imaginary part is close to 0 and the result of binarization of the imaginary part is not constant. The first non-constant area and the second non-constant area are determined independently of each other. Accordingly, the first non-constant area and the second non-constant area may be different.

In order to determine the non-constant area of the transformed iris image, a range of pixel values corresponding to the non-constant area is set. For example, when the threshold value is 0.5, a pixel having a pixel value within a range of -0.5 to 0.5 is included in the non-constant area. That is, when the absolute value of the pixel value is smaller than the threshold value, the pixel is included in the non-constant area.

If the pixels of the iris image are represented by real numbers, only one threshold is needed. However, when a pixel of the iris image is expressed as a complex number, a first threshold value for a real number and a second threshold value for an imaginary number may be required. However, according to an embodiment, the first threshold value and the second threshold value may be set to the same value.

When the pixel of the iris image is expressed as a complex number and the absolute value of the real part of the pixel is less than the first threshold value, the pixel may be included in the first non-constant area. In addition, when the absolute value of the imaginary part of the pixel is less than the second threshold value, the pixel may be included in the second non-constant area. For example, when the pixel value is 0.5-0.6j, the first threshold value is 0.6, and the second threshold value is 0.5, the absolute value of the real part of the pixel is 0.5 and the first threshold value is 0.6, so that the pixel is the first It corresponds to an unspecified area. However, since the absolute value of the imaginary part of the pixel is 0.6 and the second threshold value is 0.5, the pixel does not correspond to the second non-constant area.

9 shows a graph 900 for explaining the non-constant region in relation to the iris feature vector. As described above, when the real part or the imaginary part of a pixel expressed by a complex number is close to 0, it is determined that the binarization result of the pixel is not constant. Region A 910 includes pixels whose absolute value of the real part is less than the first threshold value 912 . Region B 920 includes pixels whose absolute value of the imaginary part is less than the second threshold value 922 . The first non-constant area is determined according to the position of the pixel corresponding to the area A 910 , and the second non-constant area is determined according to the position of the pixel in the area B 920 . Since the pixel corresponding to the region C 930 is not blocked according to the first threshold 912 and the second threshold 922 , it may be used for iris matching.

The threshold may be adaptively determined to the transformed iris image. Also, the threshold value may be a predetermined fixed value regardless of the transformed iris image. When a pixel of the iris image is expressed as a complex number, both the first threshold value and the second threshold value may be adaptively determined in the transformed iris image. Likewise, both the first threshold value and the second threshold value may be predetermined fixed values. According to an embodiment, only the first threshold value is adaptively determined for the iris image, the second threshold value may be a predetermined fixed value, and vice versa.

The threshold value may be determined according to a blocking quota indicating a ratio of pixels to be blocked. The minimum number of pixels to be blocked is determined according to the pixels of the transformed iris image and the blocking quota. Pixels are included in the non-constant area in the order in which the absolute values of the pixel values are smaller by the determined minimum quantity.

For example, if the blocking quota is 20% and the transformed iris image has 100 pixels, then at least 20 pixels must be blocked. Accordingly, 20 pixels are selected in the order of the smallest absolute values of the pixel values, and the largest value among the absolute values of the 20 pixels is determined as the threshold value. In addition, it is determined that all pixels having an absolute value greater than or less than the threshold value correspond to the non-constant area. According to an embodiment, a person skilled in the art may easily change the threshold value setting method so that only pixels having an absolute value smaller than the threshold value correspond to the non-constant area.

10 shows a specific example of determining a threshold value according to a blocking quota. Referring to FIG. 10 , pixels are arranged in the order of decreasing absolute values. If the blocking quota is 7% and there are 100 pixels in the transformed iris image, it is determined that at least 7 pixels 1010 are blocked. Accordingly, 0.2, which is the absolute value of the pixel 1012 having the 7th smallest absolute value, is determined as the threshold value. In addition, ten pixels 1020 having an absolute value equal to or smaller than 0.2 are determined as pixels included in the non-constant area.

The blocking quota may be predetermined through statistical experiments. If the blocking quota is lowered, the area used for iris matching may increase, thereby reducing the misapproval rate, but instead, the non-regular area excluded from iris matching may decrease, thereby increasing the misrejection rate. Conversely, if the blocking quota is increased, the non-regular area excluded from iris matching may increase, thereby reducing the false rejection rate, but instead, the area used for iris matching may decrease and thus the false approval rate may increase. Therefore, an appropriate blocking quota should be used to set the threshold.

The false-rejection rate mentioned above refers to the number of negative verification results that occur for a case that is positively positive, such as a case where the eyes of the same person are compared. The false acceptance rate mentioned above refers to the number of positive validation results that occur for a case that is clearly negative, such as a case where the eyes of others are compared. The false rejection rate and false acceptance rate should be kept low for robustness and accuracy of iris matching.

The allowable false rejection rate and false acceptance rate in iris recognition may be determined differently according to iris recognition accuracy required according to an iris recognition application field. Therefore, the blocking quota can be determined according to the allowable false rejection rate and false acceptance rate.

When the pixel of the iris image is expressed as a complex number, the first blocking quota applied to setting the first threshold value for the real part of the pixel and the second blocking quoting applied to setting the second threshold value for the imaginary part of the pixel are separately can be set. The first blocking quota and the second blocking quota may be set identically, or may be set to different values according to experiments.

A second mask is generated by transforming the first mask according to the determined non-constant area. As previously described, a first mask is created to block non-iris object regions from iris matching from the iris image. In addition, the second mask is generated to additionally block an irregular region including a non-constant bit in the iris object region from iris matching. An example of generating a second mask by transforming a first mask from FIG. 11 is described.

Whether each pixel of the iris image in the first mask 1110 belongs to a non-iris object region is expressed as 0 or 1. The pixel corresponding to the element determined to be 0 is not used in iris matching. Conversely, the pixel corresponding to the element determined as 1 is used for iris matching.

In the first mask 1110 , an element corresponding to pixels belonging to the non-iris object region 1112 is determined to be 0 . In addition, an element corresponding to pixels belonging to the iris object region 1114 is determined to be 1. Accordingly, as the first mask 1110 is applied to the iris image, a transformed iris image in which the non-iris object region 1112 is blocked is obtained.

The image map 1120 represents gray intensity of pixels included in the transformed iris image. Pixels belonging to the iris object region 1114 may be represented by gray intensity. The threshold value is determined according to the gray intensity of pixels belonging to the iris object region 1114 and a preset blocking quota. 11 , as an example, the threshold value is determined to be 0.2. Among pixels of the image map, a pixel having an absolute value of gray intensity less than a threshold value of 0.2 is determined to be included in the non-constant area 1122 .

The second mask 1130 is generated by changing an element belonging to the non-constant region 1122 of the first mask 1110 from 1 to 0. Accordingly, in the second mask 1130 , an element corresponding to a pixel included in the non-iris object area 1112 or the non-constant area 1122 is determined to be 0 . In addition, an element corresponding to a pixel not included in the non-iris object area 1112 or the non-constant area 1122 is determined to be 1. Accordingly, when the second mask 1130 is applied to the transformed iris image, a second transformed iris image in which the non-constant region 1122 is additionally blocked may be obtained. In addition, when the second mask 1130 is applied to an iris code to be described later, a converted iris code in which a code corresponding to the non-constant area 1122 is blocked may be generated.

In FIG. 11 , only a process of generating the second mask is illustrated when a pixel value is expressed as a real number for convenience of description. Accordingly, when a pixel value is expressed as a complex number, a second mask for the real part and a second mask for the imaginary part may be separately generated. Also, one second mask may include both information related to the blocking of the real part and the imaginary part.

In step S104, an iris code is obtained by quantizing elements included in the iris image.

The iris code may be obtained by quantizing pixel values represented by complex numbers of the iris image. As described above, binarization, which is a type of quantization, may be applied to the complex value of each pixel generated by the Gabor filter or the like. As a result of the binarization, only phase information of a complex number corresponding to pixels may be extracted as a feature of the iris pattern. The complex number of phase information is expressed as an iris feature vector, and the iris code is determined as a set of iris feature vectors of pixels.

The iris code may be obtained by binarizing pixel values represented by complex numbers of the iris image into 2 bits. One bit among the two bits may be generated by binarizing a real value of a complex value. In this specification, a bit whose real value is binarized is defined as a real binarized bit. The other 1 bit among the 2 bits may be generated by binarizing an imaginary value of a complex value. In this specification, a bit whose imaginary value is binarized is defined as an imaginary binarized bit.

As another example, the iris code may be obtained by binarizing pixel values represented by complex numbers of the transformed iris image into 2 bits. As another example, when a pixel value of the iris image is expressed as a real number, the iris code may be obtained by binarizing pixel values expressed as a real number of the iris image or the transformed iris image into 2 bits.

In step S105, a transformed iris code in which codes corresponding to the non-iris object region and the non-constant region are blocked is obtained by applying the second mask to the iris code.

The converted iris code may be obtained by separately blocking the real binarization bits and the imaginary binarization bits of the iris code according to the second mask. For example, when the real binarized bit is not a non-constant bit and the imaginary binarized bit is a non-constant bit, only the imaginary binarized bit is blocked, and the real binarized bit can be used for iris matching.

Like the first mask, the second mask may be expressed in a bit form. If each pixel of the iris image is expressed as a complex number including a real part and an imaginary part, whether or not a pixel of the iris image is blocked may be expressed by 2 bits. As with the first mask, a bit marked 0 means that the corresponding pixel or portion of the corresponding pixel is blocked from iris matching, and a bit marked 1 means that the corresponding pixel or portion of the corresponding pixel is not blocked from iris matching.

12 shows an example of applying the second masks 1210 and 1212 to the iris code 1200 . The iris code 1200 includes codes corresponding to gray intensity of pixels included in the iris image. Each code is represented by 2 bits, of which the right bit is a real binarized bit, and the left bit is an imaginary binarized bit.

To apply the second mask 1210 to the iris code 1200 , the iris code 1200 may be divided into a real iris code 1202 and an imaginary iris code 1204 . Real part iris code 1202 contains only real binarized bits, and imaginary part iris code 1204 contains only imaginary binarized bits.

A second mask 1210 for the real part is applied to the real part iris code 1202 . The real binarized bit of the pixel corresponding to the element with a value of 0 in the second mask 1210 for the real part is blocked in iris matching.

Similarly, a second mask 1212 for the real part is applied for the imaginary part iris code 1204 . The imaginary binarized bit of the pixel corresponding to the element in which the value of the second mask 1212 for the imaginary part is zero is blocked in iris matching.

The result of blocking by the second mask 1210 for the real part and the second mask 1212 for the imaginary part is merged to obtain a transformed iris code 1220 .

The blocking according to the second mask 1210 for the real part and the blocking according to the second mask 1212 for the imaginary part may be performed independently of each other. Therefore, only one unblocked bit of the code corresponding to the same pixel can be used for iris matching.

Code a 1230 having a value of '10' is blocked by both the second mask 1210 for the real part and the second mask 1212 for the imaginary part. Accordingly, the code a 1230 is expressed as '--', which means that both the real binarization bit and the imaginary binarization bit are blocked in the transform iris code 1220 .

Code b 1232 having a value of '10' is not blocked by both the second mask 1210 for the real part and the second mask 1212 for the imaginary part. Accordingly, the code b 1232 is expressed as '10', which is the same value as the iris code 1200 in the converted iris code 1220 .

The code c 1234 having a value of '01' is not blocked by the second mask 1210 for the real part, but is blocked by the second mask 1212 for the imaginary part. Therefore, the code c (1234) is expressed as '0-', which means that only the imaginary binarized bits are blocked in the transform iris code 1220.

Code d 1236 having a value of '01' is not blocked by the second mask 1212 for the imaginary part, but is blocked by the second mask 1210 for the real part. Therefore, the code d 1236 is expressed as '-1', which means that only real binarized bits are blocked in the transform iris code 1220.

In FIG. 12 , a method of blocking the code by separating the real part and the imaginary part of the iris code and the second mask has been described. However, the second mask may be applied to the iris code without the separation.

In step S106, the user is recognized by matching the reference iris code stored in advance by the user with the converted iris code. By matching the converted iris code in which the non-constant bits of the non-constant region is removed from the iris code and the reference iris code stored in advance in the user's internal or external iris code database, whether the iris code matches or whether the iris code is authentic or not is checked do. Once the match is successful or authentic, the user will be able to perform functions that require different levels of security, from social networking services to mobile banking applications, for example. Conversely, if the match fails or the authenticity is not verified, the user is denied access to the security-required function. The aforementioned pre-storing of the reference iris code may be performed in advance using the function of the iris recognition apparatus described herein.

The application fields of the proposed iris recognition method are iris-based biometric authentication for border control, technology using the iris as a biometric passport, computer login, technology using the iris as a biometric password, wireless device-based authentication technology, ATM or mobile Technology to securely access bank accounts or banking applications with applications, ticketless travel, authentication of rights to services, premises access control (home, office, laboratory, refrigerator, etc.), driver's license, other personal credentials qualifications, forensics field, birth certificate , tracking the ignition and unlocking of vehicles of missing or being pursued, anti-theft devices, anti-terrorism devices (e.g. airport security screening), financial transaction (e-commerce, finance) security, Internet security, access control to confidential information, keys , card, PIN or "biometric key encryption" (a stable key in an unstable template) used as a password. For example, a blocking quota for iris recognition in a banking application may be higher than a quota for iris recognition in a social network application. The blocking quota may be determined differently according to an accuracy of iris recognition required according to an iris recognition application field.

13 illustrates a user recognition apparatus 1300 using an iris according to an embodiment of the present invention. The user recognition apparatus 1300 includes a mask generator 1310 , an iris code acquirer 1320 , and an iris recognizer 1330 . In FIG. 13 , the encoding order mask generator 1310 , the iris code acquirer 1320 , and the iris recognizer 1330 are expressed as separate structural units, but according to an embodiment, the mask generator 1310 and the iris code acquisition part The unit 1320 and the iris recognition unit 1330 may be combined to implement the same unit.

In FIG. 13 , the mask generator 1310 , the iris code acquirer 1320 , and the iris recognizer 1330 are expressed as structural units located in one device, but the mask generator 1310 and the iris code acquirer 1320 are and devices responsible for the respective functions of the iris recognition unit 1330 are not necessarily physically adjacent to each other. Accordingly, according to an embodiment, the mask generator 1310 , the iris code acquirer 1320 , and the iris recognizer 1330 may be dispersed.

The mask generator 1310 , the iris code acquirer 1320 , and the iris recognizer 1330 may be implemented by one processor according to an embodiment. Also, according to an embodiment, it may be implemented by a plurality of processors.

The mask generator 1310 may generate a first mask that blocks the non-iris object region of the iris image. In addition, the mask generator 1310 may generate a transformed iris image in which the non-iris object region is blocked according to the first mask.

The mask generator 1310 may adaptively transform the first mask according to properties of the transformed iris image to generate a second mask that additionally blocks a non-constant region in which a quantization result of the transformed iris image is not constant. The mask generator 1310 may obtain a complex value representing a gray intensity of a pixel of the transformed iris image. The mask generator 1310 may determine a threshold value for determining non-constant bits according to the complex value and the blocking quota. The threshold value may be determined for the real part and the imaginary part of the complex value, respectively. Similarly, the blocking quota for determining the threshold may be preset for the real part and the imaginary part of the complex value, respectively.

The mask generator 1310 may determine whether to block the real part of the pixel from iris matching by comparing the absolute value of the real part of the pixel expressed as a complex number with a threshold value. Similarly, the mask generator 1310 may determine whether to block the imaginary part of the pixel from iris matching by comparing the absolute value of the imaginary part of the pixel expressed as a complex number with a threshold value. The mask generator 1310 may convert the second mask for blocking non-constant bits according to the threshold value.

When a pixel is expressed as a real number instead of a complex number, the mask generator 1310 may determine whether to block the pixel from iris matching by comparing an absolute value of the pixel expressed as a real number with a threshold value. Since there is no imaginary part in the pixel, the mask generator 1310 does not generate a mask for the imaginary part.

The iris code acquisition unit 1320 may acquire the iris code by quantizing pixels included in the iris image. The iris code obtainer 1320 may obtain the iris code by binarizing complex values corresponding to pixels of the iris image.

The iris code obtaining unit 1320 may obtain a converted iris code in which codes corresponding to the non-iris object area and the non-constant area are blocked by applying the second mask to the iris code. The iris code obtaining unit 1320 may obtain the converted iris code by separately performing a process of blocking real binarized bits and imaginary binarized bits of the iris code.

The iris recognition unit 1330 may recognize the user by matching the reference iris code stored in advance by the user with the converted iris code.

Also, at least one of the mask generator 1310 , the iris code acquirer 1320 , and the iris recognizer 1330 may be implemented as a software module. When at least one of the mask generator 1310 , the iris code acquirer 1320 , and the iris recognizer 1330 is implemented as a software module (or a program module including an instruction), the software module is a computer It may be stored in a readable non-transitory computer readable medium that can be read by a user. Also, in this case, at least one software module may be provided by an operating system (OS) or may be provided by a predetermined application. Alternatively, a part of the at least one software module may be provided by an operating system (OS), and the other part may be provided by a predetermined application.

The user recognition apparatus 1300 of FIG. 13 may perform each function and step related to iris recognition introduced in FIG. 1 .

14 illustrates a user device 1400 including a user recognition device according to an embodiment of the present invention. The user device 1400 can include a processor 1401 , a display 1402 , an infrared camera 1403 , a memory 1404 , and a keyboard 1405 .

The user device 1400 according to various embodiments of the present disclosure may include a smart phone, a tablet PC, a mobile phone, a video phone, an e-book reader, a desktop PC, a laptop PC, a PDA, a PMP, an MP3 player, a mobile medical device, a camera, or wearable devices (eg, head mounted displays (HMDs), such as electronic glasses, electronic clothing, electronic bracelets, electronic necklaces, electronic accessories, electronic tattoos, or smart watches).

The processor 1401 is configured to capture images with the camera 1403 , process the images according to the methods described herein, and store information in the memory 1404 . The processor 1401 may perform the functions of each of the components of FIG. 13 . Although the processor 1401 is represented as a single processor in FIG. 14 , it may be a plurality of processors according to an embodiment.

Display 1402 is configured to display information to a user. For example, the display may display an image, a user interface for capturing the image, iris matching results, and any other necessary information. In one embodiment, the display may be touch sensitive.

Camera 1403 is equipped with infrared illumination and is configured to perform an image capture process as directed by processor 1401 . The camera may have other types of light sources. It should be understood that the method of the present invention may be modified to use other types of light in the frame capturing step.

Memory 1404 is configured to store information. For example, the memory may store captured images, processed images, additional information about the images (eg iris codes, masks, reference iris codes, etc.).

The keyboard 1405 is used by the user to control the device. For example, a keyboard may be used to control the image capture process. The keyboard is not limited to a physical keyboard, but may be a virtual keyboard used in a touch-sensitive display.

It should be understood that the above description of the user equipment is exemplary, and the user equipment may also include other hardware, software, or firmware components in addition to or in place of the components 1401 - 1405 described above. And, the user device described above is configured to perform one or more of the steps included in any one of the methods described herein. Also, the user may generate the reference iris code in advance by using the user device. The user device may capture a facial image to extract an eye image of the user, process the eye image by at least some of the steps of the disclosed method, and store the processed eye image in memory for use in a subsequent iris matching process. .

It should be understood that the embodiments of the present specification may be implemented in real numbers in addition to or instead of the above-described complex numbers. In the above embodiment, one real number representing the gray intensity of a pixel of the iris image is binarized into 1 bit, not 2 bits applied to complex numbers as described above. Also, the iris code is a bit representation of the iris image, and the bit representation is obtained in an encoding process. A pair of bits generated from a pixel value represented by a complex number corresponds to a single point in the original image (according to the quantization of the complex number). Also, other techniques of feature extraction and encoding may be possible. For example, a Local Binary Pattern (LBP) transform may be used. The LBP transform transforms the iris image into a matrix of integers (8 bits or 16 bits depending on the selected type of LBP).

In the context of the present application, converting one real value or a complex value into one or two discrete values (bits), respectively, can be understood under the concept of quantization. Also in connection with the present application, converting a numerical value of image intensity into a bit form (iris code) may be understood under the concept of encoding. The encoding may include all steps from feature extraction to the step regarding the final form of the code to be stored in the memory.

The above description of the embodiments of the present invention is illustrative, and modifications of the construction and implementation of the present invention are within the scope of the content described herein. For example, while an embodiment of the present invention is described with respect to FIGS. 1-14 , such description is exemplary. Although the invention has been described in terms of specific structural features or methodological acts, the invention defined in the appended claims is not necessarily limited to the specific configurations or acts described. Rather, the specific configurations and acts described above are disclosed as example forms of implementing the claimed invention. Further, the present invention is not limited by the order of the method steps, and the order can be modified without creative efforts by those skilled in the art to which the present invention relates. Some or all of the method steps may be performed sequentially or concurrently.

Some embodiments may also be implemented in the form of a recording medium including instructions executable by a computer, such as program modules to be executed by a computer. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transport mechanism, and includes any information delivery media. Also, some embodiments may be implemented as a computer program or computer program product including instructions executable by a computer.

Claims (15)

In the user recognition method using the iris,
generating a first mask that blocks a non-iris object area of the iris image;
generating a transformed iris image in which the non-iris object region is blocked according to the first mask;
obtaining complex values corresponding to pixels of the transformed iris image;
separating real and imaginary values of the complex values to generate a real-valued array and an imaginary-valued array;
obtaining a first threshold applied to the real-valued array according to a first blocking quota applied to the real-valued array;
obtaining a second threshold applied to the array of imaginary values according to a second blocking quota applied to the array of imaginary values;
a second mask for additionally blocking an inconsistent area in which a quantization result of the transformed iris image is not constant by transforming the first mask according to the first threshold value and the second threshold value; generating;
obtaining an iris code by quantizing pixels included in the iris image;
applying the second mask to the iris code to obtain a transformed iris code in which the non-iris object region and a portion corresponding to the non-constant region are blocked; and
and recognizing the user by matching the reference iris code stored in advance by the user with the converted iris code. According to claim 1,
obtaining an eye image from the face image;
dividing an iris image expressed according to an angular coordinate system from the eye image; and
and normalizing the iris image to be expressed according to a Cartesian coordinate system. According to claim 1,
The step of obtaining the iris code is
obtaining the iris code by binarizing complex values corresponding to pixels of the iris image;
Each of the complex values is binarized into 2 bits, and 1 bit of the 2 bits is a real binarized bit generated by binarizing the real value of the complex value, and the other 1 bit of the 2 bits is generated by binarizing the imaginary value of the complex value. A user recognition method using an iris, characterized in that it is an imaginary binarized bit. 4. The method of claim 3,
The step of obtaining the converted iris code comprises:
and obtaining the converted iris code by separately blocking real binarized bits and imaginary binarized bits of the iris code according to the second mask. 4. The method of claim 3,
The step of obtaining the iris code is
and binarizing complex values corresponding to pixels of the iris image by applying a Gabor filter to the iris image. 6. The method of claim 5,
Each complex value corresponding to the pixels of the iris image is determined from the gray intensity corresponding to the pixel of the iris image. delete delete According to claim 1,
The first blocking quarter represents a predetermined ratio of real values not used for iris matching among real values included in the real value array,
and the second blocking quarter represents a predetermined ratio of imaginary values not used for iris matching among imaginary values included in the imaginary value array. According to claim 1,
Obtaining the first threshold value comprises:
acquiring a first threshold value according to the absolute value magnitude of real values included in the real value array and the first blocking quota;
Transforming the first mask comprises:
and converting the first mask so that a real part of a pixel corresponding to a real value having an absolute value smaller than the first blocking quarter is not used in iris matching. According to claim 1,
Obtaining the second threshold value comprises:
acquiring a second threshold according to the absolute value magnitude of imaginary values included in the imaginary value array and the second blocking quota;
Transforming the second mask comprises:
and converting the first mask so that an imaginary part of a pixel corresponding to an imaginary value having an absolute value smaller than the second blocking quarter is not used in iris matching. delete delete In the user recognition device using the iris,
create a first mask that blocks the non-iris object region of the iris image;
generating a transformed iris image in which the non-iris object region is blocked according to the first mask;
A first block applied to the real value array by obtaining complex values corresponding to pixels of the transformed iris image, separating real and imaginary values of the complex values, to generate a real value array and an imaginary value array obtaining a first threshold applied to the array of real values according to a quota, and obtaining a second threshold applied to the array of imaginary values according to a second blocking quota applied to the array of imaginary values, wherein the first threshold a mask generator that transforms the first mask according to a value and the second threshold to generate a second mask that additionally blocks a non-constant region in which a quantization result of the transformed iris image is not constant;
obtaining an iris code by quantizing pixels included in the iris image;
an iris code acquisition unit configured to obtain a converted iris code in which a portion corresponding to the non-iris object region and the non-regular region is blocked by applying the second mask to the iris code; and
and an iris recognition unit configured to recognize the user by matching the reference iris code stored in advance by the user with the converted iris code. A computer program product comprising a computer-readable storage medium, the storage medium comprising:
A computer program product storing instructions for performing each step of the user recognition method using the iris of claim 1 .

Images