KR101423153B1 - Invariant radial iris segmentation - Google Patents

Abstract

Methods and computer products are provided for identifying objects by biometric analysis of the eye. First, an iris image to be recognized is obtained. Texture enhancement may be performed on the image as desired but is not required. Next, the iris image is radially segmented into a selected number of radial segments, for example 200, each radial segment representing 1.8 [deg.] Of the iris scan. After segmentation, each radial segment is analyzed and peak and valleys of color intensity are detected in the iris radial segment. These detected peaks and valleys are converted mathematically into a set of data used to construct the template. The template represents the scanned and analyzed iris of the object and is constructed from data transformed from the radial segment, respectively. After building, this template is stored in the database and used for matching purposes if the object has already been registered in the database.

Description

{INVARIANT RADIAL IRIS SEGMENTATION}

[Related Application]

This application is related to U.S. Provisional Patent Application No. 11 / 043,366, filed January 26, 2005 and entitled " AID Polar Based Segmentation Approach ". The disclosure of related documents is entirely incorporated herein by reference.

[Declaration of Government Rights]

The invention was made with government support under Contract No. F10801 EE5.2. The government has certain rights to the invention.

[TECHNICAL FIELD]

The present invention relates to biometrics, and more particularly to an improved approach to radial iris segmentation.

Biometrics is the study of automated methods that uniquely identify humans according to one or more unique physical or behavioral characteristics. In the field of IT, biometrics refers to the technique of measuring and analyzing human physical characteristics for authentication purposes. Examples of physical characteristics are fingerprints, eye retina and iris, facial patterns and hand measurements.

A major concern of conventional biometric systems is that the lack of precise acquisition of biometric data or differences in operating conditions can easily miss personal characteristics that distinguish humans from other humans. Iris recognition has been considered as a method of achieving low error and high success in restoring biometric data. However, iris scanning and image processing are expensive and time consuming. Fingerprints, facial patterns and hand measurements have provided cheaper and faster solutions.

Over the past several years, iris recognition has matured enough to compete economically with other biometric methods. However, discrepancies in the recognition condition of the iris image cause rejection of the valid object and approval of the impersonator, especially when performed under environmental conditions where scanning is not controlled.

In contrast, iris recognition under controlled conditions proved to be very effective. This is true because the iris recognition system relies on features that are more distinct than other biometric techniques, such as facial patterns, hand measurements, and thus provide a reliable solution by providing a much more distinct biometric data set.

Although prototype systems and techniques were proposed in the early 1980s, independent iris recognition systems have been developed since the 1990s. The concepts found in this study were then implemented in real devices. The overall approach is based on the conversion of raw iris images into numeric codes that can be easily implemented. The robustness of this approach and the following alternative approach is based on thorough iris segmentation. Iris segmentation is the process of locating and separating iris from other parts of the eye. Calculation of the iris feature requires a high quality segmentation process that focuses on the iris of the subject and extracts the border appropriately. This acquisition process is sensitive to acquisition conditions and proved to be a very interesting problem. Conventional systems attempt to maximize segmentation precision by limiting operating conditions. The illumination level, the position of the scanned eye, and the ambient temperature. This restriction may enable more precise iris recognition, but is not practical in real-time manipulation.

Significant progress has been made to minimize these problems; However, this evolution has been largely based on the original method, ie circular / elliptical outer segmentation, which proved to be a problem under uncontrolled conditions. Another study introduces a concept that has a similar problem of intensifying segmentation under uncontrolled conditions that compete with the methods discussed above.

Thus, a method for providing iris recognition technology that is well suited for iris-at-a-distance applications, i. E. Systems that utilize unconstrained conditions, while still providing accurate and real-time results in accordance with the collected biometric data Lt; / RTI >

In accordance with the principles of the present invention, a new feature extraction technique is provided with a new encoding scheme that yields an improved biometric algorithm. This new extraction technique is based on simplified polar segmentation (POSE). The new encoding scheme uses the new extraction technique to launch the actual local iris feature using processing with a low computational load.

The encoding scheme does not rely on precise segmentation of the outer boundary of the iris region, which is necessary in the prior art. Instead, it relies on the identification of peaks and valleys in the iris (i.e., noticeable change points in the color strength of the iris). Advantageously, regardless of the selected filter, the encoding scheme does not depend on the exact point of occurrence of the peak detected in the iris, but instead depends on the magnitude of the detected peak relative to the first peak referenced. Since this algorithm does not depend on the exact location of the pattern peak / valley, precise segmentation to the outer boundary of the iris is not necessary, which eliminates the need for normalization processing.

The overall function of the present invention can be summarized as follows. First, the iris is preprocessed and then localized using an enhanced radial POSE segmentation and enhanced segmentation process based on the POSE method referred to herein. During the segmentation process, all unclear parts (i.e., pupil, eyelid, eyelash, sclera and other nonessential parts of the eye) are excluded from the analysis if they are different at the inner boundary of the iris. Illumination correction and contrast enhancement are handled to compensate for differences in image illumination and reflection conditions. The captured iris image is unwrapped into a plurality of radial segments, and each segment is analyzed to produce a one-dimensional dataset representing peak and / or valley data of the segment. The peak and / or valley data is one dimensional in the sense that the peaks and / or valleys are aligned along their outwardly directed straight line from the center of the iris along their position. In one embodiment, the iris image is unwrapped with a one-dimensional pole representation for an iris signature where data for only one peak per radial segment is stored. In one embodiment, the size of the outermost peak from the pupil-iris boundary per segment is stored. In another embodiment, the magnitude of the maximum peak of the segment is stored. In another embodiment, data for a plurality of peaks and / or valleys per radial segment is stored. In this embodiment, each peak and / or valley is recorded with a one-bit value representing the magnitude relative to other peaks and / or valleys in the segment, such as a peak / valley immediately preceding the one-dimensional direction. The data for all radial segments is connected to a template representing the data for the entire iris scan. The template is compared with the stored template to find the coincidence.

Figure 1 shows a scanned iris image according to the prior art.

2A shows a scanned iris image using the principles of the present invention.

FIG. 2B shows a scanned iris image of FIG. 2A mapped to a one-dimensional iris map.

Figure 3 shows a flow chart illustrating one embodiment of the present invention.

Figure 4A illustrates the mapping of the iris segmentation process in accordance with the principles of the present invention.

Figure 4B illustrates an enhanced iris scan mapping in accordance with the principles of the present invention.

Figure 5A illustrates a first encoding scheme in accordance with the principles of the present invention.

Figure 5B illustrates a second encoding scheme in accordance with the principles of the present invention.

A primary concern of conventional biometric systems is that the lack of precise acquisition of biometric data or differences in operating conditions can easily obscure personal characteristics that distinguish humans from other humans. Over the past few years, iris recognition has matured in that it is comparable to more common biometric measures such as fingerprints. However, the discrepancy in the acquisition conditions of the iris image causes the valid object to be rejected and the impersonator to be approved, especially under an operating environment in which the light is not closely controlled. In contrast, under controlled conditions, iris recognition has proven to be very efficient. This is because the iris recognition system relies on features that are more distinct than other common biometric methods, which provide reliability by providing more pronounced biometrics.

Figure 1 shows a scanned eye image with boundaries identified in accordance with conventional segmentation techniques. Here, the iris 105 is defined by the external iris boundary 110. However, the outer iris boundary 110 is blocked by the eyelid at 107, and the actual boundary can not be determined. The system must guess the missing portion of the outer iris boundary 110. [ Computing the iris feature requires a high-quality segmentation process that focuses on the iris of the subject and extracts the boundary appropriately. This process is sensitive to acquisition conditions and proved to be a very interesting problem (especially for non-cooperative objects captured remotely). Attention is directed to addressing conventional segmentation problems by limiting operating conditions such as controlled illumination and eye position of the object, but such an approach is not always practical.

The main failure cause of the prior art is that the system focuses on the outer boundary of the iris to normalize iris scanning to allow for uniform matching. Because of the many factors, including the iris, which can make the external iris boundaries unclear, and the light color iris, which can make the distinction between the eyelashes and the sclera difficult, the outer boundaries may not be precisely mapped and as a result, Resulting in inaccurate segmentation and negatively affecting the rest of the biometric recognition process. Also, when applied to uncontrolled conditions, such a segmentation technique causes many errors. Such conditions include objects that may not be directly aligned with the captured object or imaging device in various ranges from the acquiring device.

2A shows an image of a scanned eye similar to FIG. In this figure, the principles of the present invention are applied. This method is based on simplified pole segmentation (POSE), a newer encoding scheme that does not rely on precise segmentation of the outer boundary of the iris. A detailed description of POSE can be found in the aforementioned co-pending U. S. Patent Application Serial No. 11 / 043,366 entitled " A 1D Polar Based Segmentation Approach ". The present invention utilizes enhanced POSE technology. This enhanced POSE technique or invariant radial POSE focuses on detecting peak discontinuities in the intensity of color between the pupil and the sclera within the defined peak and valleys of the iris, i. E., The defined radial segment of the iris. In other words, the peak is the point at which the color intensity at any one side (in the selected direction) of the point is weaker than the color intensity at that point (also, to prevent all of the weak discontinuities from being registered as record peaks The discontinuity point exceeds a predetermined threshold value). Similarly, the valley is the point at which the color intensity at any one side of the point in the selected direction is greater than the color intensity at that point (with the same limitations).

This technique is referred to as one dimension because iris data collected per radial segment has only one signal dimension instead of collecting two-dimensional image data per radial segment as in the prior art. This eliminates the need to guess the unclear outer boundary of the iris and the need to calculate the exact parameters of the circle, ellipse, or any other shape needed to guess the missing part of the outer boundary.

The iris 205 is scanned using an invariant radial POSE process. Rather than focusing on the outer boundaries of the iris as in the process of FIG. 1, the invariant radial POSE process locates and identifies the peaks and valleys present in the scanned iris and generates an iris map. FIG. 2A helps illustrate one form of iris map that can represent peak and / or valley data in an iris scan. In Figure 2A, data for only one peak per radial segment is stored. In order to construct an iris map according to this embodiment of the present invention, the iris is first segmented into a set of radial segments, for example 200 segments. Thus, each segment represents a 1.8-degree piece of the entire 360-degree scan of the iris. After each of the 200 segments is analyzed, data for one characteristic peak is stored in the segment. In the embodiment illustrated in Figures 2A and 2B, the representative selected peak in each radial segment is the peak 210 at the outermost point from the pupil-iris boundary. In other embodiments, the selected peak may be the largest peak (except for the peak at the pupil-iris boundary), the steepest peak, or the innermost peak. If the criterion is the outermost peak, the nearer the iris-scleral boundary, the less peak and valley become less clear as a reference for identifying the object and thus less reliable, so that the pupil- It is preferable to use the outer peak.

Instead, the data corresponding to the valley may be recorded instead of the peak. In fact, the recorded data may not be a peak or valley and may be any other readily identifiable color or contrast characteristic. The distance from the center of the pupil is stored even if any peak or valley (or other characteristic) is selected as representative. In a preferred embodiment, the radial distance is reported as a relative value relative to the radial distance of the reference peak from the center of the pupil. In this way, there is no need to normalize the iris scan to compensate for variations in iris due to environmental conditions (e.g., pupil dilation, ambient light). In a preferred embodiment of the present invention, the reference peak is the peak at the pupil-iris boundary in the segment, which is usually, but not always, the maximum peak within that segment.

2B shows a scanned iris mapped to a one-dimensional iris map. To construct such an iris map, the iris is first segmented into a predetermined number of radial segments, for example, 200 segments, each segment representing 1.8 degrees during a full 360 degree scan of the iris. After each of the 200 segments is analyzed, a reference peak is selected in each segment, and the reference peak is the peak at the radial segment pupil-iris boundary analyzed (generally, but not always, the largest peak in that segment). The iris is unfolded to produce the graph shown in Figure 2B, where each point 215 represents the aforementioned relative radial distance of the corresponding peak for each radial segment.

For visual purposes, the conversion of peak and valley data to the graph shown in FIG. 2B can be thought of as "unwrapping" the iris with respect to the normal of the pupil-iris boundary. For example, the pupil-iris boundary is essentially a circular boundary. As shown in FIG. 2, assume that the boundary is a string and unwrapping the string with a straight line having a reference peak from each radial segment as discontinuity point 215,

The following description is merely exemplary in nature to enable those of ordinary skill in the art to instinctively understand the process. Those of ordinary skill in the relevant art will appreciate that converting peak and valley information to a one-dimensional dataset is in fact a simple mathematical transformation.

Irrespective of the conditions such as illumination and temperature (which affect the expansion or contraction of the pupil diameter), the one-dimensional iris representation does not change with respect to the relative position of the reference peaks in each angular segment, It can lead to downward movement. Pupil expansion and other factors can affect the absolute position of the peak or valley (i.e., the actual distance from the pupil boundary), but affect the relative position of the peak and valleys in the iris relative to the reference peak (or valley) Do not.

Figure 4A shows the formation of another more robust representation of scanned iris data in which data is recorded for a number of peaks that are not one characteristic peak per radial segment. The center of the pupil is denoted by the cross (405). The horizontal or x-axis represents the radial distance from the pupil-iris boundary (i.e., perpendicular to the pupil-iris boundary), and the vertical or y-axis represents the derivative of color intensity. The peak at the pupil-iris boundary is denoted by 411. All other peaks and valleys in the segment are graphically displayed relative to the reference peak so that data normalization is not required.

It should be noted that each radial segment generally has a width of several pixels at the pupil boundary 410 and wider as the distance from the pupil-iris boundary increases. Thus, in order to generate the one-dimensional data shown in the graph of FIG. 4A, the color intensity derivative data represented by the y-axis must be averaged and interpolated over the width of the segment. This representation of the interpolated data is shown at line 415, where each significant data peak is denoted by reference numeral 420.

Figure 4B shows yet another embodiment. The graph 425 shows the graph representation of the iris as shown in FIG. 4A. According to the embodiment of Figure 4A, in the feature extraction step, preferably each individual peak is recorded separately for a reference peak. However, also, to focus only on the peaks, enhancement curves 430 are removed from the one-dimensional iris representation. The enhancement curve 430 is a graph component that can be removed without nourishing the magnitude of each peak relative to the next peak producing a normalized data set that only focuses on the magnitude of the relative peak. Using standard wavelet analysis known to those of ordinary skill in the art, the enhancement curve can be calculated as an approximate component (DC component) for the decomposition of the graph of FIG. 4A. Once the enhancement curve is removed, each peak becomes a segmented graph 435 represented by a point 437 on the graph 435. However, with the elimination of the enhancement curve, the graph 425 is normalized with respect to peak occurrence. As will be described in greater detail below, in at least one embodiment of the present invention, peak data can be encoded very efficiently by encoding each peak for an adjacent peak using fewer bits, such as one or two bits per peak There will be. Thus, the removed enhancement curve simplifies the process while retaining all the necessary information.

FIG. 3 shows a flowchart showing an embodiment of the present invention.

In step 305, a preprocessing step is performed. Pretreatment can be formal in nature. At this stage, texture enhancement is performed on the scanned image. Unclear parts of images such as pupils, eyelids, eyelashes, nettings and other non-essential parts of the eye are excluded from the analysis. To reduce the side effects of external illumination and grayscale variations and other artifacts (e.g., color contact lenses), the system pre-processes the image using a local radial texture pattern (LRTP). However, it should be noted that the texture enhancement is not essential to the operation of the system.

The images are available from Y. Du, R. Ives, D. Etter, T. Welch, C.-I. It is similar to that proposed by CHang, "A one-dimensional approach for iris identification", EE Dept, US Naval Academy, Annapolis, MD, 2004, but is preprocessed using the improved local radial texture pattern.

here,

I (x, y) = the color intensity of the pixel located at the two-dimensional coordinate x, y;

ω = a curve that determines the neighboring point of pixel x, y; And

Area of A = ω (number of pixels)

to be.

This LRTP method is different from the LRTP method because it avoids discontinuity by block analysis employed in the above-mentioned reference document while maintaining the approximation of the actual average using the window average instead. The average of each window of the small block contains a rough guess of the background illumination and is therefore subtracted from the actual value of the intensity as shown in the previous equation.

In step 310, an invariant radial POSE segmentation process is performed. This method differs from traditional techniques because it does not require iris segmentation at the outer boundary of the iris, the iris-sclera boundary.

In particular, we first roughly determine the iris center in the original image, then refine the center estimate and extract the edge of the pupil. Techniques for centering the pupil are disclosed in the aforementioned U.S. Patent Application No. 11 / 043,366, which is incorporated by reference and need not be further described. Techniques for locating pupil-iris boundaries are also disclosed in the aforementioned patent application and need not be described further.

Once the pupil edge is known, the segmentation process is initiated. A radial scan of the iris is performed in a radial segment, for example 200 segments of 1.8 degrees each.

After segmentation and scanning at step 310, the process proceeds to step 315. [ In step 315, the actual feature extraction occurs based on the segmented image obtained in step 310. The feature extraction process may be performed in accordance with one of the three embodiments for detecting variations in the graph representation of the iris, for example, without depending on the absolute position of the occurrence of the variation described above with respect to Figures 2A and 2B, 4A and 4B, respectively . In particular, the absolute position varies as a function of the natural expansion and contraction of the human iris when exposed to variations in environmental lighting conditions. Thus, the feature extraction process relies on detecting the relative variation of peaks and valleys in size and their relative position rather than focusing on its absolute magnitude or position. The main advantage of this method is that it does not require a normalization procedure of iris scans to compensate for variations in iris due to environmental conditions. The iris normalization procedure is very important for conventional iris recognition technology.

Next, at step 320, the peak data according to the result expressed in the graph 435 is encoded into a template that is encoded to be subsequently compared efficiently with a stored template of iris data for a known person. Two different encodings are discussed below with respect to Figures 5A and 5B, respectively. These two are only examples, and are not intended to limit the scope of the invention.

5A and 5B each show the encoding of a peak / valley dataset for one radial segment of the iris scanned according to two embodiments of the present invention. As will be described in more detail below, each template includes a plurality of such data sets, and the number of such assemblies in the template is equal to the number of radial segments. Thus, for example, if each segment is 1.8 [deg.], Each template contains 200 such data sets.

5A shows a first encoding scheme focused on the magnitude of the relative peak relative to the magnitude of the immediately preceding peak. FIG. 5A shows encoding of peak data for a single radial segment and represents a dataset for that segment. Each data set contains IxK bits, where K is the number of peaks per radial segment for recording a peak, and I is the number of bits used to encode each peak. K may be any suitable number and should be chosen close to the number of typical peaks expected in the radial segment. In Fig. 5A, K = 8.

In this encoding scheme, the first I bit of all data sets represents the selected reference peak (pupil-iris boundary), and is set to a first value, for example 11, when I = 2. When moving from left to right within the dataset, the bits represent peaks radially outward from the pupil-iris boundary (i.e., the x-axis in Figures 2B, 4A, and 4B, which represent the distance from the pupil- . If the magnitude of the peak in the graph such as graph 435 is greater than the magnitude of the previous peak, the bit representing the peak is set to 11. Otherwise, the bit is set to a second value, for example, 00. Thus, in this example it is guaranteed that the reference peak has the largest magnitude in that segment and is therefore always larger than the next peak, thus ensuring that the second I bit is essentially 00. Therefore, in this encoding scheme, the first four bits of each data set will always be equal to 1100, so it is irrelevant to the match and will not be considered during the match. If the radial segment does not have at least a K-peak, the end of the data set is filled with one or more third bit sets having a third value of, for example, 10 or 01 that will ultimately be masked in the matching step 325. If the radial segment has more than K-peaks, only the K-peak closest to the pupil-iris boundary is encoded.

Thus, with respect to the iris segment shown in the first left graph of Fig. 5A, the sequence representing the peak / valley information for this iris segment is 1100110011001010. In particular, the first two bits represent the size of the reference peak 501 and are always 11, the second two bits represent the size of the first peak 503 in the segment and are always guaranteed to be 00 because they will always be smaller than the reference peak The fifth and sixth bits are 11 because the next peak 505 is greater than the previous peak 503 and the seventh and eighth bits are 00 because the next peak 507 is smaller than the previous peak 507 , The ninth and tenth bits are 00 because the next peak 511 is less than the immediately preceding peak 509 and the last four bits indicate that this segment has only five peaks (and the reference peak is the sixth peak ), It is 1010 corresponding to two unknown sets.

5A, the first two bits indicate the size of the reference peak 501 and are always 11, and the next two bits indicate the size of the first peak in the segment 513, 00 and the next two bits are 00 since the next peak 515 is greater than the immediately preceding peak 513 and the next two bits indicate that the next peak 517 is greater than the immediately preceding peak 517 00, and the next two bits are 11 because the next peak 519 is larger than the immediately preceding peak 517, and the last six vigorous segments have only five peaks (including the reference peak) 101010 Then, the peak / valley information for the segment of the iris is 1100000011101010.

Figure 5B illustrates an exemplary second encoding scheme in accordance with the principles of the present invention. The second encoding scheme is also based on 2-bit quantification of the size of the filtered peaks, but the peak size is quantified to three magnitude levels: low (L), high (H) and medium (M). The low level size L is assigned a 2-bit pattern 00, and the high level size H is assigned a 2-bit pattern 11. To solve the quantization error, the level is constructed in such a way as to allow only one bit of tolerance to shift from one adjacent quantization level to the next. By this restriction, the scheme has two combinations for expressing the medium level, namely Mi = 10 and Mr = 01. Mi denotes the case where the valley to the left of the corresponding peak is lower than the valley to the right of the peak, and Mr denotes the peak whose valley to the right is lower than the valley to the left. For example, peak 520 may be labeled Mi because the valley to the left is lower than the valley to the right. Conversely, the peak 521 may be marked Mr because the valley to the left is higher than the valley to the right. However, the two result values for the size of the medium level are treated the same in the matching process. Also, if there is not enough peak to fill the dataset, the bits are used to complete the bit-vector to a predefined vector length, for example 10 bits. Although not shown in the figure, bits corresponding to unknown peaks can be identified by any suitable means, such as adding a flag to the end of the data set representing the number of bits corresponding to an unknown peak. Alternatively, the level can be encoded with 3 bit quantization to provide an additional bit combination to indicate that it is not known. In addition, one value, e.g., 100, may be assigned for the medium level, leaving a two bit combination of 01 to indicate that it is unknown. The bits that are not known are masked during matching as described below. Similarly, when the number of peaks in the radial segment exceeds the number of peaks required to fill the dataset, the peaks farthest from the pupil are excluded.

Next, at step 325, the template is constructed by concatenating all the data sets corresponding to all the radial segments in the iris scan. Thus, for example, if there are 200 radial segments and the number of bits used for each data set in the encoding scheme representing the detected peaks is 16 bits, then all the encoded binary strings are encoded into a template of 16 x 200 = 3400 bits .

If the data is encoded, the process continues to step 330. [ This process determines whether the scanned iris template matches the stored iris template by comparing similarities between the corresponding bit-templates. The weighted Hamming distance can be used as a recognition measure to perform a bitwise comparison. The comparison algorithm may include a noise mask to mask the bits indicating that only significant bits may be used to calculate the information measurement distance (e.g., Hamming distance). The algorithm reports the value according to the comparison. The higher value reflects the lower similarity in the template. Thus, the lowest value is considered to be the best matching score in the two templates.

After the information measurements for the two templates have been computed to resolve the inconsistencies and image misalignment due to rotation, one template is moved to the left and right bits (along the axis of rotation), and a plurality of information measurement distance values It is calculated from the movement. Such movement in the bit direction in the angular direction corresponds to the rotation of the circular iris region by the angle resolution unit. Only the lowest value from the calculated information measurement distance is considered to be the best matching score in the two templates.

The weighting mechanism may be used in conjunction with the above-described matching. The bits representing the nearest peaks to the pupil region (pupil boundary) are the most reliable and / or clear data points, and can be weighted higher because they represent more precise data. All bits representing unknowns are weighted with zero for matching whether they are in the matching target template or the stored template. This can be done using any suitable technique. In one embodiment, when two templates are compared, the bit position corresponding to the bit representing the unknown of one or two templates is always filled with the bits matching the corresponding bit of the other template.

Although the embodiments described above depend on the direction and the analysis of the peaks in the iris, this is shown as an example only. Other embodiments may depend on the detection of the valley of the iris or any other distinct feature in the iris.

It will be apparent to those skilled in the art that the procedures, procedures, and / or steps of the invention described herein may be performed by a programmed computer device execution software designed to perform them. Moreover, such processes, procedures, and / or steps may be performed in the form of circuitry including, but not limited to, application specific integrated circuits (ASICs), logic circuits, and state machines.

Although specific embodiments of the invention have been described, it will be apparent to those skilled in the art that various alternatives, modifications, and improvements may be readily apparent to those skilled in the art. These alternatives, modifications, and improvements, obvious to those skilled in the art, are not intended to be set forth herein but are intended to be part of the description and are intended to be within the spirit and scope of the present invention. Accordingly, the foregoing description is by way of example only and is not restrictive. The invention is limited only as defined in the following claims and their equivalents.

Claims (38)

(1) acquiring an object iris image; (2) segmenting the iris image radially into a plurality of radial segments; (3) for each radial segment, determining data for a predetermined one-dimensional feature in the segment relative to a reference value for the feature in the image; And (4) constructing a template for the object including a collection of the data for each of the radial segments; Lt; / RTI > Wherein the feature comprises at least one of a peak and a valley of color intensity, The reference value is a peak or valley of color intensity selected to represent each of the radial segments, Wherein the data comprises at least one of a relative size of at least one of the peaks and valleys and a relative position of at least one of the peaks and valleys along a direction radially outwardly from the central pupil of the subject. A method for identifying an object by biometric analysis. The method according to claim 1, The step (2) (2.1) determining the center of the pupil of the object; (2,2) determining pupil-iris boundaries in the image; And (2.3) segmenting the iris into a plurality of radial segments having the same angular size; The method comprising the steps of: identifying an object by an iris biometric analysis of an eye; (1) acquiring an object iris image; (2) segmenting the iris image radially into a plurality of radial segments; (3) for each radial segment, determining data for a predetermined one-dimensional feature in the segment relative to a reference value for the feature in the image; And (4) constructing a template for the object including a collection of the data for each of the radial segments; Lt; / RTI > The step (2) (2.1) determining the center of the pupil of the object; (2,2) determining pupil-iris boundaries in the image; And (2.3) segmenting the iris into a plurality of radial segments having the same angular size; / RTI > Wherein the feature comprises a peak of color intensity and the data includes a distance from the pupil-iris boundary of one of the peaks farthest from the pupil-iris boundary within a predetermined distance of the pupil-iris boundary Wherein the object is identified by a biometric analysis of the iris of the eye. (1) acquiring an object iris image; (2) segmenting the iris image radially into a plurality of radial segments; (3) for each radial segment, determining data for a predetermined one-dimensional feature in the segment relative to a reference value for the feature in the image; And (4) constructing a template for the object including a collection of the data for each of the radial segments; Lt; / RTI > The step (2) (2.1) determining the center of the pupil of the object; (2,2) determining pupil-iris boundaries in the image; And (2.3) segmenting the iris into a plurality of radial segments having the same angular size; / RTI > Wherein the one-dimensional data comprises a distance from a pupil-iris boundary of a largest one of peaks of color intensity in the radial segment. . delete The method according to claim 1, Wherein the relative size is determined by interpolated data for a width of each of the radial segments. The method according to claim 1, Wherein the size is determined by data averaged over the width of each of the radial segments. ≪ RTI ID = 0.0 > 31. < / RTI > 3. The method of claim 2, Wherein the reference value is one of the peaks corresponding to the intensity of the color at the pupil-iris boundary. 9. The method of claim 8, Wherein said step (3) comprises removing a decomposition curve from said peaks and valleys. 3. The method of claim 2, (5) encoding the data into a data set; The method further comprising the steps of: identifying the subject by the biometric analysis of the iris of the eye. 11. The method of claim 10, In the step (5) Wherein a predetermined first number of bits is used to represent data for each peak in the radial segment, each of the data sets having a predetermined second number of peaks corresponding to a predetermined number of peaks that can be encoded in the data set Bits, If the number of peaks in the radial segment is less than the predetermined number of peaks that can be encoded, filling the data set with bits representing an unknown peak, And encoding a subset of the peaks in the radial segment if the number of peaks in the radial segment is greater than the predetermined number of peaks that can be encoded, A method of identifying an object by biometric analysis of an iris of the eye. delete 12. The method of claim 11, Wherein the subset of peaks comprises the peak closest to the pupil of the subject in the radial segment. ≪ RTI ID = 0.0 > 31. < / RTI > 12. The method of claim 11, Wherein the subset of peaks comprises the largest peak detected in the radial segment. ≪ RTI ID = 0.0 > 11. < / RTI > 11. The method of claim 10, (6) comparing the template of the subject to at least one stored template to determine if the template of the subject matches the at least one stored template; The method further comprising the steps of: identifying the subject by the biometric analysis of the iris of the eye. 16. The method of claim 15, Wherein step (6) comprises weighting each of the encoded data sets such that the bits corresponding to the peaks closer to the pupil of the object are weighted more than the bits farther from the pupil of the object Wherein the object is identified by a biometric analysis of the iris of the eye. 11. The method of claim 10, The step (6) Determining an order of the peaks in the radial segment; Assigning a first value to each of the peaks having a magnitude greater than the previous peak in the sequence; Assigning a second value to each of the peaks having a magnitude smaller than the previous peak in the sequence; And Placing the values in the data set according to the order; The method comprising the steps of: identifying an object by an iris biometric analysis of an eye; 18. The method of claim 17, Wherein the reference value is one of the peaks corresponding to a color intensity at the pupil-iris boundary, Wherein the peak is aligned with the distance from the pupil of the subject. ≪ RTI ID = 0.0 > 11. < / RTI > delete delete The method according to claim 1, The step (1) Performing a texture enhancement on the image to pre-process the image, and excluding a portion of the image that makes the iris unclear. ≪ RTI ID = 0.0 > How to. A first computer executable instruction for obtaining an image of an iris of a subject; A second computer execute instruction to radially segment the image for the iris into a plurality of radial segments; Third computer-executable instructions for determining data for a predetermined one-dimensional feature within the segment relative to a reference value for the feature within the image; And A fourth computer-executed instruction for building and storing a template for the object including the set of data for each of the radial segments; Lt; / RTI > Wherein the feature comprises at least one of a peak and a valley of color intensity, The reference value is a peak or valley of color intensity selected to represent each of the radial segments, Wherein the data comprises at least one of a size of at least one of the peaks and valleys and a relative position of at least one of the peaks and valleys along a direction radially outward from the central pupil of the subject by biometric analysis of the iris of the eye A computer program product recorded on a computer readable medium for identifying an object. 23. The method of claim 22, Wherein the second computer execute instruction comprises: Determining a center of the pupil of the object; Determining a pupil-iris boundary in the image; And Segmenting the iris into a plurality of radial segments having the same angular size; A computer program product recorded on a computer readable medium for identifying an object by biometric analysis of an iris of an eye. A first computer executable instruction for obtaining an image of an iris of a subject; A second computer execute instruction to radially segment the image for the iris into a plurality of radial segments; Third computer-executable instructions for determining data for a predetermined one-dimensional feature within the segment relative to a reference value for the feature within the image; And A fourth computer-executed instruction for building and storing a template for the object including the set of data for each of the radial segments; Lt; / RTI > Wherein the second computer execute instruction comprises: Determining a center of the pupil of the object; Determining a pupil-iris boundary in the image; And Segmenting the iris into a plurality of radial segments having the same angular size; / RTI > Characterized in that the characteristic is a peak of color intensity and the data comprises a distance from the pupil-iris boundary of one of the peaks farthest from the pupil-iris boundary within a predetermined distance of the pupil-iris boundary A computer program product recorded on a computer readable medium for identifying an object by biometric analysis of an iris of the eye. A first computer executable instruction for obtaining an image of an iris of a subject; A second computer execute instruction to radially segment the image for the iris into a plurality of radial segments; Third computer-executable instructions for determining data for a predetermined one-dimensional feature within the segment relative to a reference value for the feature within the image; And A fourth computer-executed instruction for building and storing a template for the object including the set of data for each of the radial segments; Lt; / RTI > Wherein the second computer execute instruction comprises: Determining a center of the pupil of the object; Determining a pupil-iris boundary in the image; And Segmenting the iris into a plurality of radial segments having the same angular size; / RTI > Wherein the one-dimensional data comprises a distance from a pupil-iris boundary of a largest one of peaks of hue intensity in the radial segment to identify an object by an iris biometric analysis of the eye A computer program product recorded on a computer readable medium. delete 23. The method of claim 22, Wherein the relative size is determined by interpolated data for a width of each of the radial segments. ≪ Desc / Clms Page number 20 > 23. The method of claim 22, Characterized in that the size is determined by data averaged over the width of each of the radial segments. ≪ RTI ID = 0.0 > < / RTI > 24. The method of claim 23, Wherein the reference value is one of the peaks corresponding to the intensity of the color at the pupil-iris boundary. ≪ Desc / Clms Page number 21 > 30. The method of claim 29, Wherein the third computer-executable instructions comprise instructions for removing a decomposition curve from the peaks and valleys. ≪ Desc / Clms Page number 21 > 31. The method of claim 30, Further comprising fifth computer instructions for encoding the data into a data set, wherein the fifth computer instructions encode the data into a data set. 32. The method of claim 31, In the fifth computer instruction, Wherein a predetermined first number of bits is used to represent data for each peak in the radial segment, each of the data sets having a predetermined second number of peaks corresponding to a predetermined number of peaks that can be encoded in the data set Bits, If the number of peaks in the radial segment is less than the predetermined number of peaks that can be encoded, filling the data set with bits representing an unknown peak, And encoding a subset of the peaks in the radial segment if the number of peaks in the radial segment is greater than the predetermined number of peaks that can be encoded, A computer program product recorded on a computer readable medium for identifying an object by biometric analysis of an iris of an eye. 33. The method of claim 32, Characterized in that the subset of peaks comprises the peak closest to the pupil of the subject in the radial segment. ≪ RTI ID = 0.0 > 15. < / RTI & product. 33. The method of claim 32, Characterized in that the subset of peaks comprises the largest peak detected in the radial segment. ≪ Desc / Clms Page number 13 > 33. The method of claim 32, A sixth computer execute instruction to compare the template of the subject to at least one stored template to determine if the template of the subject matches the at least one stored template; Further comprising a biometric analysis of the iris of the eye. ≪ Desc / Clms Page number 13 > 36. The method of claim 35, Wherein the sixth computer-executable instructions include weighting each of the encoded data sets such that the bits corresponding to the peaks closer to the pupil of the object are weighted more than the bits farther away from the pupil of the object A computer program product recorded on a computer readable medium for identifying an object by biometric analysis of an iris of an eye. 36. The method of claim 35, The sixth computer execute instruction comprises: Determining an order of the peaks in the radial segment; Assigning a first value to each of the peaks having a magnitude greater than the previous peak in the sequence; Assigning a second value to each of the peaks having a magnitude smaller than the previous peak in the sequence; And Placing the values in the data set according to the order; A computer program product recorded on a computer readable medium for identifying an object by biometric analysis of an iris of an eye. 39. The method of claim 37, Wherein the reference value is one of the peaks corresponding to a color intensity at the pupil-iris boundary, Wherein the peak is aligned with the distance from the pupil of the subject. ≪ RTI ID = 0.0 > 11. < / RTI >

Images