APPARATUS AND METHOD OF EVALUATING IMAGE QUALITY
FOR REALTIME IRIS RECOGNITION, AND STORAGE MEDIA
HAVING PROGRAM THEREOF
Technical Field
The present invention relates to an apparatus and a method of evaluating image quality for real-time iris recognition, and a storage medium having a program thereof. More specifically, the present invention relates to an apparatus and a method of evaluating image quality for real-time iris recognition, and a storage medium having a program thereof which are capable of detecting only a suitable image in an iris recognition in real time, and providing an image in which a priority of image quality is determined so as to confirm an image of excellent quality first.
Background Art Conventionally, an individual personal password or an individual personal identification number is widely used for uniquely identifying a living being. The living being includes humans such as an individual person and animals. However, such a conventional method may not always stable and accurately identify the individual person due to theft and loss. In addition, due to an inverse function thereof, the conventional method causes unfavorable side effects covering the world at large. Biological recognition has been in the spotlight as a method to compensate for such a disadvantage of the conventional method.
Biological recognition is a method that uniquely identifies the individual person based on the individual person's physical (biological) characteristics and active characteristics. The physical characteristics include a fingerprint, a face, an iris, and a palm print. The active characteristics include a signature and a voice.
Iris recognition is best among various biological recognition based on unity, constancy, and stability aspects. The iris recognition has a very low wrong recognition rate and is applicable to a field that requires a high level of security. The iris is entirely formed before a person becomes three years old. As long as the person does not receive a specific external wound, the iris remains unchanged.
A conventional iris recognition method does not use a filtering which judges' compatibility of an image in real time with respect to the image inputted from a camera. Consequently, an unsuitable image lowers the recognition rate of a total system. Furthermore, the conventional iris recognition method requires an individual person (referred to as "user") to accurately operate and select a location, causing user inconvenience. An image input having low quality is an obstacle to the automation of an iris recognition system. Although a filtering operation is added, to calculate it takes much time. The filtering operation is used in only conditional circumstance.
Disclosure of the Invention
Therefore, the present invention has been made in view of the above- mentioned problems, and it is an object of the present invention to provide an apparatus and a method of evaluating image quality for real-time iris recognition, and a storage medium having a program thereof which are capable of detecting only a suitable image in an iris recognition in real time in order to quickly and accurately identify an individual person.
It is another object of the present invention to provide an apparatus and a method of evaluating image quality for a real-time iris recognition, and a storage medium having a program thereof which are capable of providing a priority of image quality so as to confirm an image of excellent quality first.
According to an aspect of the present invention, there is provided a An apparatus of evaluating image quality for real-time iris recognition comprising: a sampling means for receiving an image of an eye captured by a pick-up unit, for dividing the receiving image of an eye into pixel regions of predetermined units, and for obtaining an average of the pixel regions in order to sample the image of an eye; an absolute evaluation means for detecting a pupil from the sampled image of an eye from the sampling means, for absolutely evaluating image quality of the detected pupil to obtain an absolute evaluated value, and for eliminating an image having an absolute evaluated value less than an absolute reference value and outputting an image having an absolute evaluated value greater than the absolute reference value; and a relative evaluation means for restoring a pupil of the sampled image from the absolute evaluation means to an original size, for matching a pupil of the restored region with the photographed image of an eye, and relatively determining a priority of the image quality using an eyelid location of the sampled image.
Preferably, the absolute evaluation means includes a pupil detector for determining a threshold value of the pupil using a histogram indicating a brightness distribution of an eye image, for binarizing the sampled image based on the determined threshold value of the pupil, and vertically/horizontally projecting the binarized sampled image in order to obtain a size and a location of the pupil; an absolute evaluation section for absolutely evaluating the image quality based on the size and the location of the pupil obtained by the pupil detector; and an eliminator for eliminating an image having the absolutely evaluated value from the absolute evaluation section less than the absolute reference value, and for outputting an image having the absolutely evaluated value from the absolute evaluation section equal to or greater than the absolute reference value.
Preferably, the absolute evaluation section includes: an existential evaluating module for evaluating a size of a pupil using an existential evaluating function S
p :
where, R
mem is an average of sizes of pupils, x
radlm is an x axis radius of a pupil, y
r di
us is
a y
aχi
s radius of the pupil, and the average R
mean of sizes of pupils is a value averaging sizes of pupils in images which can be recognized; a location evaluating module for evaluating a location of a pupil using a location evaluating function S
j :
where, I
wm is a transverse length of an image, I
hejgM is a longitudinal length of the image, x
c is an x coordinate of an image center, y
c is a y coordinate of the image center, x is an x coordinate of a detected pupil center, and y is a y coordinate of the detected pupil center. a completeness evaluating module for evaluating an aspect ratio of the pupil using a completeness evaluating function S :
Ss =^^, for xradius ≥ y radius radius
Xradius
^s ~ > J0r X radius \ y radius • radius
Preferably, the relative evaluation means includes a searching region section for setting a searching region having a pupil from the sampled image outputted by the absolute evaluation means; a restoring section for restoring the searching region set by the searching region section to an original size, and for matching a pupil of the restored region with the photographed image of an
2168
5 eye; an eyelid detector for detecting a horizontal edge from the matched image from the restoring section, for horizontally projecting the detected horizontal edge, and detecting an eyelid location of the sampled image using the horizontally projected horizontal edge; and an absolute evaluation section for relatively determining a priority of the image quality according the eyelid location of the sampled image and a distance between the eyelid and the pupil.
Also, the searching region includes a region of a square, the square circumscribes the pupil. The eyelid detector detects the horizontal edge using a Sobel-Y operator. The eyelid detector detects a location in which the most horizontal components appear as the eyelid location.
There is also provided a method of evaluating image quality for real-time iris recognition comprising the steps of:
(i) receiving an image of an eye captured by a pick-up unit; (ii) dividing the receiving image of an eye into pixel regions of predetermined units, and for obtaining an average of the pixel regions in order to sample the image of an eye;
(iii) detecting a pupil from the sampled image of an eye, absolutely evaluating image quality of the detected pupil to obtain an absolute evaluated value, and eliminating an image having an absolute evaluated value less than an absolute reference value and outputting an image having an absolute evaluated value greater than the absolute reference value; and
(iv) restoring a pupil of the sampled image to an original size, matching a pupil of the restored region with the photographed image of an eye, and relatively determining a priority of the image quality using an eyelid location of the sampled image.
Preferably, step (iii) includes the steps of: (a) determining a threshold value of the pupil using a histogram indicating a brightness distribution of an eye image;
6
(b) binarizing the sampled image based on the determined threshold value of the pupil; (c) vertically/horizontally projecting the binarized sampled image in order to obtain a size and a location of the pupil; (d) absolutely evaluating the image quality based on the size and the location of the pupil obtained in step (c); and (e) eliminating an image having the absolutely evaluated value less than the absolute reference value, and outputting an image having the absolutely evaluated value equal to or greater than the absolute reference value.
Preferably, step (d) includes the steps of:
(dl) evaluating a size of a pupil using an existential evaluating function S
p
where, R
mean is an average of sizes of pupils, x
rad„„ is an x axis radius of a pupil, y,
adms is a y axis radius of the pupil, and the average R
mem of sizes of pupils is a value averaging sizes of pupils in images which can be recognized; (d2) evaluating a location of a pupil using a location evaluating function s, =\~ * width
X +- * height ~ y „ y
c where, I
mM is a transverse length of an image, I
heιght is a longitudinal length of the image, x
c is an x coordinate of an image center, y
c is a y coordinate of the image center, x is an x coordinate of a detected pupil center, and y is a y coordinate of the detected pupil center.
(d3) evaluating an aspect ratio of the pupil using a completeness evaluating function Ss :
ci _ "ladms fnr „ ■> .,
"s — ' JUI Λradnιs ~ radius
Xraduιs
X
"_■ ~ s J0r X radius \ y radius ' y radius
Preferably, step (iv) includes the steps of: (f) setting a searching region having a pupil from the sampled image outputted; (g) restoring the set searching region to an original size, and matching a pupil of the restored region with the photographed image of an eye; (h) detecting a horizontal edge from the matched image from the restoring section 234; (i) horizontally projecting the detected horizontal edge; (j) detecting an eyelid location of the sampled image using the horizontally projected horizontal edge; and (k) relatively determining a priority of the image quality according to the eyelid location of the sampled image and a distance between the eyelid and the pupil.
Preferably, step (f) sets a region of a square, the square circumscribes the pupil. Step (h) detects the horizontal edge using a Sobel-Y operator. Step (j) detects a location in which the most horizontal components appear as the eyelid location.
There is provided a storage medium having a program for evaluating image quality for real-time iris recognition comprising: a process for receiving an image of an eye captured by a pick-up unit; a process for dividing the receiving image of an eye into pixel regions of predetermined units, and for obtaining an average of the pixel regions in order to sample the image of an eye; an absolute evaluating process for detecting a pupil from the sampled image of an eye, absolutely evaluating image quality of the detected pupil to obtain an absolute evaluated value, and eliminating an image having an absolute evaluated value less than an absolute reference value and outputting an image having an absolute evaluated value greater than the absolute reference value; and a relative evaluating process for restoring a pupil of the sampled image to an original size, matching a pupil of the restored region with the photographed
image of an eye, and relatively determining a priority of the image quality using an eyelid location of the sampled image.
Preferably, the absolute evaluating process includes: (a) a process for determining a threshold value of the pupil using a histogram indicating a brightness distribution of an eye image; (b) a process for binarizing the sampled image based on the determined threshold value of the pupil; (c) a process for vertically /horizontally projecting the binarized sampled image in order to obtain a size and a location of the pupil; (d) a process absolutely evaluating the image quality based on the size and the location of the pupil; and (e) a process for eliminating an image having the absolutely evaluated value less than the absolute reference value, and for outputting an image having the absolutely evaluated value equal to or greater than the absolute reference value.
Preferably, the process (d) includes:
(dl) a process of evaluating a size of a pupil using an existential evaluating function S„ :
where, R
mem is an average of sizes of pupils, x
radius is an x axis radius of a pupil, y
r i
us is
a y axis radius of the pupil, and the average R
mean of sizes of pupils is a value averaging sizes of pupils in images which can be recognized; (d2) a process of evaluating a location of a pupil using a location evaluating function S, :
where, I
wjdth is a transverse length of an image, I
lιejglli is a longitudinal length of the image, x
c is an x coordinate of an image center, y
c is a y coordinate of the image
center, x is an x coordinate of a detected pupil center, and is a y coordinate of the detected pupil center.
(d3) a process of evaluating an aspect ratio of the pupil using a completeness evaluating function Ss : s = , JOr faά a — y radius radius
Preferably, the absolute evaluating process includes: (f) a process of setting a searching region having a pupil from the sampled image outputted; (g) a process of restoring the set searching region to an original size, and matching a pupil of the restored region with the photographed image of an eye; (h) a process of detecting a horizontal edge from the matched image; (i) a process of horizontally projecting the detected horizontal edge; (j) a process of detecting an eyelid location of the sampled image using the horizontally projected horizontal edge; and (k) a process of relatively determining a priority of the image quality according to the eyelid location of the sampled image and a distance between the eyelid and the pupil.
Preferably, the process (f) sets a square that circumscribes the pupil to the searching region. The process (h) detects the horizontal edge using a Sobel-Y operator. The process (j) detects a location in which the most horizontal components appear as the eyelid location. A preferred embodiment of the present invention includes a computer system and a computer program product programmed to execute a method according the present invention. According to the computer system, a set of command signals is stored in one or more memories. The set of command signals may be stored in a recording medium such as a CD-ROM as the computer program product.
Brief Description of the Drawings
The foregoing and other objects, features and advantages of the present invention will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which: FIG. la is a block diagram showing a configuration of an apparatus of evaluating image quality for real-time iris recognition according to an embodiment of the present invention;
FIG. lb is a block diagram showing an example of an absolute evaluation section shown in FIG. la; FIG. 2a is a flow chart that illustrates a method of evaluating image quality for real-time iris recognition according to an embodiment of the present invention;
FIG. 2b is a flow chart that illustrates step S430 in detail shown in FIG. 2a;
FIG. 2c is a flow chart that illustrates step S440 in detail shown in FIG. 2a;
FIG. 3 is a view for showing brightness characteristics by regions of an eye image;
FIGS. 4a through 4d are views for showing images unsuitable for iris recognition;
FIG. 5 is a view for showing a binarized sampled image;
FIG. 6 is a view for showing an example which horizontally/vertically projects the binarized sampled image;
FIG. 7 is a view for showing an example which restores a searching region having a pupil to an original size, and matches a pupil of the restored region with the photographed image of an eye;
FIG. 8 is a view for showing a horizontally projected horizontal edge; and FIG. 9 is a view for showing images in which priorities of quality are determined.
Best Mode for Carrying Out the Invention
Reference will now be made in detail to the preferred embodiments of the present invention.
FIG. la is a block diagram showing a configuration of an apparatus 200 of evaluating image quality for a real-time iris recognition according to an embodiment of the present invention. FIG. lb is a block diagram showing an example of an absolute evaluation section shown in FIG. la.
As shown in FIG. la, the sampling unit 210 receives an image of an eye captured by a pick-up unit 100. The pick-up unit 100 includes a camera. The sampling unit 210 divides the receiving image of an eye into pixel regions of predetermined units. The sampling unit 21 obtains an average of the pixel regions in order to sample the image of an eye.
The absolute evaluation unit 220 detects a pupil from the sampled image of an eye from the sampling section 210. The absolute evaluation section 220 absolutely evaluates image quality of the detected pupil to obtain an absolute evaluated value. The absolute evaluation unit 220 eliminates an image having an absolute evaluated value less than an absolute reference value, and outputs an image having an absolute evaluated value greater than the absolute reference value. The absolute evaluation unit 220 includes a pupil detector 222, an absolute evaluation section 224, and an eliminator 226.
The pupil detector 222 determines a threshold value of the pupil using a histogram indicating a brightness distribution of an eye image. In an embodiment of the present invention, preferably, the histogram includes an intensity histogram. The pupil detector 222 binarizes the sampled image based on the determined threshold value of the pupil. The pupil detector 222 vertically/horizontally projects the binarized sampled image in order to obtain a size and a location of the pupil.
The absolute evaluation section 224 absolutely evaluates the image quality
based on the size and the location of the pupil obtained by the pupil detector 222. As shown in FIG. lb, the absolute evaluation section 224 includes an existential evaluating module 224a, a location-evaluating module 224b, and a completeness- evaluating module 224c. A description will now be given, with reference to FIG. 6, of a process for obtaining the absolute evaluated value.
An existential evaluating module 224a of the absolute evaluation section 224 evaluates a size of a pupil using an existential evaluating function SP . The existential evaluating function SP is expressed as the following Equation 1. Equation 1
S =1- lr-R„ ,r = radius mean y radius where, R
mean is an average of sizes of pupils, x
radjus is an x axis radius of a pupil, y
ra i
us i
a y axis radius of the pupil, and the average R
mem of sizes of pupils is a value averaging the sizes of pupils in images which can be recognized. A location-evaluating module 224b of the absolute evaluation section 224 evaluates a location of a pupil using a location S, . The location evaluating function S
! is expressed as the following Equation 2. Equation 2
where, I
wjdth is a transverse length of an image, I
hejglu is a longitudinal length of the image, x
c is an x coordinate of an image center, y
c is a y coordinate of the image center, x is an x coordinate of a detected pupil center, and y is a y coordinate of the detected pupil center.
A completeness-evaluating module 224c of the absolute evaluation section 224 evaluates an aspect ratio of the pupil using a completeness evaluating function
Ss . The completeness evaluating function Ss is expressed as the following Equation 3. Equation 3
C- _ radius f >
°s ■> Jυr Λ radius — y radius radius
X Λ^ = ? jor xradjlιs yradjus . y radius
The eliminator 226 eliminates an image having the absolutely evaluated value from the absolute evaluation section 224 less than the absolute reference value, and outputs an image having the absolutely evaluated value from the absolute evaluation section 224 equal to or greater than the absolute reference value. The relative evaluation unit 230 includes a searching region section 232, a restoring section 234, an eyelid detector 236, and an absolute evaluation section 238.
The searching region section 232 sets a searching region having a pupil from the sampled image outputted by the absolute evaluation unit 220. The restoring section 234 restores the searching region set by the searching region section 232 to an original size. The restoring section 234 matches a pupil of the restored region with the photographed image of an eye. The eyelid detector 236 detects a horizontal edge from the matched image from the restoring section 234. The eyelid detector 236 horizontally projects the detected horizontal edge. The eyelid detector 236 detects an eyelid location of the sampled image using the horizontally projected horizontal edge. The absolute evaluation section 238 relatively determines a priority of the image quality according to the eyelid location of the sampled image and a distance between the eyelid and the pupil.
The priority of the image quality determined by the relative evaluation unit 230 and an original image passed through an absolute evaluating process are outputted to an iris recognition system 300. The iris recognition system 300
recognizes an iris indicative of an image in order from the highest priority to lowest priority.
Hereinafter, a method of evaluating image quality for real-time iris recognition according to an embodiment of the present invention will be described with reference to FIGS. 2a through 2c. FIG. 2a is a flow chart that illustrates a method of evaluating image quality for a real-time iris recognition according to an embodiment of the present invention.
A sampling unit 210 receives a photographed image of an eye from the outside (step S410). The sampling unit 210 samples an average of the image by a predetermined pixel region unit (step S420). The absolute evaluation unit 220 detects a pupil from the sampled image of an eye from the sampling unit 210. The absolute evaluation unit 220 eliminates an image having an absolute evaluated value less than an absolute reference value (step S430). The relative evaluation unit 230 restores a pupil of the sampled image from the absolute evaluation unit 220 to an original size. The relative evaluation unit 230 matches a pupil of the restored region with the photographed image of an eye. The relative evaluation unit 230 relatively determines a priority of the image quality using an eyelid location of the sampled image (step S440).
As shown in FIG. 2b, step S430 includes the following steps: a step S431 of determining a threshold value of the pupil using a histogram indicating a brightness distribution of an eye image; a step S432 of binarizing the sampled image based on the determined threshold value of the pupil; steps S433 and S434 of vertically/horizontally projecting the binarized sampled image in order to obtain a size and a location of the pupil; a step S435 of absolutely evaluating the image quality based on the size and the location of the pupil obtained in steps S433 and
S434; and a step S436 of eliminating an image having the absolutely evaluated value less than the absolute reference value, and outputting an image having the
absolutely evaluated value equal to or greater than the absolute reference value.
Step S435 includes the following steps: step (dl) (not shown) of evaluating a size of a pupil using the existential evaluating function SP expressed as the following Equation 1; step (d2) (not shown) of evaluating a location of the pupil using the location evaluating function S, expressed as the following Equation 2; and step (d3) (not shown) of evaluating the completeness evaluating function Ss expressed as the following Equation 3.
As shown in FIG. 2c, step S440 includes the following steps: a step S441 of setting a searching region having a pupil from the sampled image outputted; a step S442 of restoring the set searching region to an original size, and matching a pupil of the restored region with the photographed image of an eye; a step S443 of detecting a horizontal edge from the matched image from the restoring section 234; a step S444 of horizontally projecting the detected horizontal edge; a step S445 of detecting an eyelid location of the sampled image using the horizontally projected horizontal edge; and a step S446 of relatively determining a priority of the image quality according to the eyelid location of the sampled image and a distance between the eyelid and the pupil.
Step S441 sets a region of a square. The square circumscribes the pupil. Step S443 detects the horizontal edge using a Sobel-Y operator. Step S445 detects a location in which the most horizontal components appear as the eyelid location.
In the present invention constructed as above, a description will now be given, with reference to FIGS, la through 8, of a method for detecting only a suitable image in an iris recognition in real time, and determining and providing a priority of the image quality to the iris recognition system 300 by the image quality evaluator 200 in order to quickly and accurately identify an individual person.
The sampling unit 210 receives an image of an eye captured by a pick-up unit 100 (step S410). The sampling unit 210 divides the receiving image of an eye
into pixel regions of predetermined units, and obtains an average of the pixel regions in order to sample the image of an eye (step S420). For example, the image of an eye is divided into four pixel regions in transverse and longitudinal directions, respectively. An average of four pixel regions is taken to obtain the result that
1 diminishes a total size of the image of an eye by — . Doing that will reduce a
16 processing time.
Thereafter, the absolute evaluation unit 220 detects a pupil from the sampled image of an eye from the sampling section 210, and eliminates an image having an absolute evaluated value less than an absolute reference value, and outputs an image having an absolute evaluated value greater than the absolute reference value (step S430). Hereinafter, step S430 will be explained with reference to FIG. 3. FIG. 3 is a view for showing brightness characteristics by regions of an eye image. In a histogram of the eye image that can be recognized, the eye image is divided into three parts having a brightness of a pupil region, a brightness of an iris region, and a brightness of a sclera. The brightness of the pupil region is located at the lowest part in the histogram and has a feature that is distinguished from the two other parts.
However, an image unsuitable in the iris recognition exists among the images as shown in FIGS. 4a through 4d. When comparing the images of FIGS. 4a through 4d with the image of FIG. 3, a peak of the pupil region does not suitably appear in each histogram of FIGS. 4a through 4d. FIGS. 4a through 4d are views for showing images unsuitable for iris recognition. FIG. 4a shows an image when the user blinks the user's eye. FIG. 4b shows an image when dust enters into the user's eye. FIG. 4c shows an image when an eyelash or an eyelid irritates the pupil. FIG. 4d shows an image when a center of the pupil is quite removed from a center of the image.
In order to improve efficiency and recognition rate of an iris recognition
process, the images unsuitable for iris recognition should be excluded prior to the iris recognition.
In order to use such characteristics, the pupil detector 222 determines a threshold value of the pupil using a histogram indicating a brightness distribution of an eye image (step S431). The pupil detector 222 searches a minimal point between two peaks PI and P2 shown in FIG. 3 and determines the threshold value of the pupil based on the searched minimal point. The reason is that an image in which a peak appears clear is perfectly distinguished from other parts according to a threshold value determined in step S431, whereas an image in which a peak does not appear clear is perfectly distinguished from the other parts. The pupil detector 222 binarizes the sampled image based on the determined threshold value of the pupil. FIG. 5 shows the binarized sampled image. Thereafter, the pupil detector 222 vertically/horizontally projects the binarized sampled image (step S433) and obtains a size and a location of the pupil (step S434). FIG. 6 is a view for showing an example which horizontally/vertically proj ects the binarized sampled image.
The absolute evaluation section 224 absolutely evaluates the image quality based on the size and the location of the pupil obtained by the pupil detector 222 (step S435). An existential evaluating module 224a of the absolute evaluation section 224 evaluates a size of a pupil using an existential evaluating function Sp . A location-evaluating module 224b of the absolute evaluation section 224 evaluates a location of a pupil using a location S, . A completeness-evaluating module 224c of the absolute evaluation section 224 evaluates an aspect ratio of the pupil using a completeness evaluating function S^ .
The eliminator 226 sums the size of a pupil evaluated by the existential evaluating module 224a, the location of a pupil evaluated by the location evaluating module 224b, and the aspect ratio of the pupil evaluated by the completeness evaluating module 224c. The eliminator 226 eliminates an image having the
summed value less than the absolute reference value (step S436). The absolute reference value is referred to as "utilization fidelity of the image". The strength of the utilization fidelity of the image changes by adjusting the absolute reference value. In other words, since an image of good quality is applicable to an image for a registration, by adjusting the absolute reference value to a great value, a high level of recognition accuracy is achieved. In a real recognition step, the absolute reference value is adjusted to a small value. When adjusting the absolute reference value to the small value, an image of bad quality can be used for an image for iris recognition. The conventional method photographs the iris many times due to a rejection of an image adoption. However, since it is unnecessary to photograph the iris many times, the present invention is convenient for the user.
The relative evaluation unit 230 restores a pupil of the sampled image from the absolute evaluation unit 220 to an original size, matches a pupil of the restored region with the photographed image of an eye, and relatively determines a priority of the image quality using an eyelid location of the sampled image.
The searching region section 232 sets a searching region having a pupil from the sampled image outputted by the absolute evaluation unit 220 (step S441).
The restoring section 234 restores the searching region set by the searching region section 232 to an original size. The restoring section 234 matches a pupil of the restored region with the photographed image of an eye (step S442). The captured image of an eye is received from the pick-up unit 100.
FIG. 7 is a view for showing an example that restores a searching region having a pupil to an original size, and matches a pupil of the restored region with the photographed image of an eye. Such a process is performed to extract exact location information. The searching region is preferably a square that circumscribes the pupil.
02 002168
19
For example, when multiplying x direction start and end points, y direction start and end points of the pupil extracted from the sampled image by four, respectively, the location of the pupil is matched to a location of an original image.
The reason to multiply the points by four is that the average of the pixel regions is taken with four pixel region units.
The eyelid detector 236 detects a horizontal edge from the matched image from the restoring section 234 (step S443, and horizontally projects the detected horizontal edge (step S444). Although there are a plurality of methods to detect the horizontal edge, the eyelid detector 236 preferably detects the horizontal edge using a Sobel-Y operator. The eyelid detector 236 detects an eyelid location of the sampled image using the horizontally projected horizontal edge (step S445). FIG. 8 is a view for showing a horizontally projected horizontal edge.
The absolute evaluation section 238 relatively determines a priority of the image quality according to the eyelid location of the sampled image and a distance between the eyelid and the pupil (step S446). FIG. 9 is a view for showing images in which priorities of quality are determined. Since the further a distance between the eyelid and the pupil is, then the greater a region of an iris which can be recognized increases, and as a result the priority of the image quality is determined so that an image having a further distance is selected. The priority of image quality determined by the absolute evaluation section 238 is outputted to the iris recognition system 300 together with the image executing the absolute evaluating process.
Industrial Applicability
As can be seen from the foregoing, the present invention detects only a suitable image in iris recognition in real time, and provides the detected image to an iris recognition system so that the iris recognition system quickly and accurately identifies an individual person. Also, the present invention automates the iris
recognition system. The present invention detects a pupil using a histogram without being restricted by an external environment. The present invention increases efficiency of preprocessing using the detected pupil information. The present invention provides a determined priority of image quality to an iris recognition system using a location of an eyelid so that the iris recognition system recognizes an image of good quality first. A manager's instruction to adopt an image of good quality is not required in a real security system. Since images which can be recognized are adopted but images which can not be recognized are automatically eliminated among a series of continuously captured images for a predetermined time, a user easily acquires an image. Furthermore, the present invention maximizes the user's convenience.
While this invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiment and the drawings, but to the contrary, it is intended to cover various modifications and variations within the spirit and scope of the appended claims.