JP2008046813A - Condition determination method, registering device, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enhance an authentication precision. <P>SOLUTION: This condition determination method includes: calculating correlation statistics SCRm of correlation values CRm<SB>1</SB>to CRm<SB>(n-1)</SB>in n images IMm<SB>1</SB>to IMm<SB>n</SB>and pattern statistics SPQm of blood vessel pattern quantities PQm<SB>1</SB>to PQm<SB>n</SB>in binary images MIMm<SB>1</SB>to MIMm<SB>n</SB>acquired as a predetermined blood vessel extraction processing result to the images in every stage LU<SB>1</SB>to LU<SB>m</SB>; and determining illumination luminance corresponding to the stable section of the distribution of both the correlation statistics SCRm and pattern statistics SPQm in the stages LU<SB>1</SB>to LU<SB>m</SB>as one of conditions suitable for a blood vessel pattern. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a condition determination method, a registration device, and a program, and is suitable for application to biometric authentication.

  Conventionally, an authentication device fixes a finger on an imaging camera, for example, and images a fingerprint on the fixed finger. Then, the authentication device registers the fingerprint pattern image obtained as a result of imaging in a memory or the like as information for identifying the captured biological image, or compares the fingerprint pattern image with the registered fingerprint pattern image. Whether or not there is is determined.

In addition to such authentication using a fingerprint pattern, authentication using a finger blood vessel pattern has also been applied to improve the authentication accuracy of whether or not the person is a registered person (see, for example, Patent Document 1). When imaging a blood vessel, near-infrared light having a characteristic that is specifically absorbed by hemoglobin is irradiated on the living body, and the blood vessel projected onto the image sensor through the living body is imaged.
JP 2003-303178 A

  However, in the living body, the imaging conditions are uniform regardless of the degree of transmission of near-infrared light and the amount of blood vessels among individuals. For example, the amount of blood vessels is small and transmission of near-infrared light is low. When a living body whose amount is inferior becomes a registrant, the registered blood vessel pattern itself is simplified, and there is a problem that authentication accuracy as to whether or not the person is a registrant falls.

  The present invention has been made in consideration of the above points, and intends to propose a condition determination method, a registration apparatus, and a program that can improve authentication accuracy.

  In order to solve such a problem, the present invention is a condition determination method, in which an imaging condition of an imaging unit and an extraction condition for extracting a biological pattern from each image signal sequentially output as an imaging result of the biological pattern in the imaging unit, Among these, the first step of fixing the exposure time other than the exposure time of the image sensor included in the imaging means and switching the exposure time to a plurality of stages, and the correlation value between images adjacent to each other on the time axis for each image signal For each stage, and a second step for obtaining the statistic of the biometric pattern amount extracted from the image signal for each stage, and the correlation value statistic and the statistic of the biometric pattern quantity at each stage. Of the distribution, at the exposure corresponding to the middle position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlap or the position near it Was acceptable to provide a third step of determining an imaging condition suitable for the biometric pattern.

  Further, the present invention is a registration apparatus, wherein an extraction unit that extracts a biological pattern from each image signal sequentially output from an imaging element included in the imaging unit, and controls the imaging unit and the extraction unit, as an imaging result of the biological pattern Control means, the control means is fixed in the imaging conditions of the imaging means and the extraction conditions of the extraction means other than the exposure time of the imaging device, the exposure time is switched to a plurality of stages, for each image signal, The statistics of correlation values between images adjacent to each other on the time axis are obtained for each stage, the statistics of the amount of biological pattern extracted from the image signal are obtained for each stage, and the statistics of correlation values at each stage and In the distribution of the statistic of the amount of biological pattern, at the middle position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlap or in the vicinity thereof The corresponding exposure time is determined as an imaging condition suitable for the biological pattern, and one or more biological patterns extracted from each image signal output from the image sensor are stored as an imaging result of the biological pattern under the imaging condition. Added to the media registration.

  Furthermore, the present invention provides a control unit that controls an imaging unit and an extraction unit that extracts a biological pattern from each image signal sequentially output from an imaging element included in the imaging unit as a biological pattern imaging result. On the other hand, among the imaging conditions of the imaging means and the extraction conditions of the extraction means, the first processing for fixing the exposure time other than the exposure time of the imaging device and switching the exposure time to a plurality of stages, and for each image signal on the time axis The second process for obtaining the statistical amount of the correlation value between the adjacent images in each step and the statistical amount of the biological pattern extracted from the image signal for each step, and the correlation value statistics in each step In the distribution of the quantity and the statistic of the biometric pattern quantity, the intermediate position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlap or close to it The third process for determining the exposure time corresponding to the position as an imaging condition suitable for the biological pattern, and 1 or 1 extracted from each image signal output from the imaging device as the imaging result of the biological pattern under the imaging condition The fourth process of registering two or more biological patterns in the storage medium is executed.

  The statistical value distribution of the correlation value roughly reflects individual differences and the imaging conditions for the individual differences, and the extraction conditions for the individual differences are added to the statistical pattern distribution of the biological pattern quantity. It is reflected in. Therefore, according to the present invention, it is better to determine the imaging condition suitable for the biological pattern by focusing on both distributions than when determining the imaging condition suitable for the biological pattern by focusing only on one distribution. An imaging condition can be determined as being suitable for the biometric pattern, and thereby it is possible to obtain a biometric pattern that can improve the authentication accuracy, and thus a condition determination method, a registration device, and a program that can improve the authentication accuracy. realizable.

  Hereinafter, an embodiment to which the present invention is applied will be described in detail with reference to the drawings.

(1) Overall Configuration of Authentication Device According to this Embodiment FIG. 1 shows the overall configuration of the authentication device 1 according to this embodiment. The authentication device 1 includes an authentication unit 2 and a plurality of blood vessel imaging units 3 _N (N = 1, 2,...).

The plurality of blood vessel imaging units _3N have the same configuration, and are mounted on, for example, a device that unlocks doors installed at a plurality of locations (hereinafter referred to as a door unlocking device).

(1-1) Configuration of Authentication Unit This authentication unit 2 provides a bus for the operation unit 11, blood vessel imaging unit 12, blood vessel extraction unit 13, flash memory 14, notification unit 15, and communication unit 16 to the control unit 10. The blood vessel imaging units 3 _{1 to} 3 _N are connected to the communication unit 16 via the communication unit (not shown) of each door unlocking device.

  The control unit 10 includes a CPU (Central Processing Unit) that controls the entire authentication unit 2, a ROM (Read Only Memory) that stores various programs and setting information, and a RAM (Random Access Memory) as a work memory of the CPU. ) And a microcomputer.

The control unit 10 is provided with an execution command COM1 of a mode for registering a blood vessel of a user to be registered (hereinafter referred to as a registrant) (hereinafter referred to as a blood vessel registration mode) via the operation unit 11. is also registrant presence determining mode (hereinafter, the authentication mode hereinafter) executes instructions COM2 of vascular imaging unit 3 _1, 3 ₂ ..., or from the door unlocking device incorporating a 3 _N It is given via the communication unit 16.

  The control unit 10 determines a mode to be executed according to the execution instructions COM1 and COM2, and executes the blood vessel registration mode or the authentication mode based on a program corresponding to the determination result.

(1-2) Configuration of Blood Vessel Imaging Unit The blood vessel imaging unit 12 is configured by connecting a near-infrared light source 22, an imaging camera 23, and a memory 24 to the drive unit 21, respectively. The memory 24 stores a value for correcting sensitivity variation for each pixel in the image sensor ID (hereinafter referred to as a correction coefficient).

  The variation in sensitivity for each pixel in the image pickup element ID is caused by the nonuniformity of the light incident aperture shape with respect to the pixel, and therefore the signal value for each pixel has a linear characteristic with respect to the amount of incident light. Therefore, the correction coefficient is the reciprocal of the sensitivity variation for each pixel with respect to the signal value for each pixel.

  In order to calculate such a correction coefficient, a subject having a uniform amount of light is imaged in advance (hereinafter referred to as calibration), and a signal value (hereinafter referred to as “this”) for each pixel of the image sensor ID obtained at this time. The average value (referred to as calibration signal value) can be obtained by dividing by the calibration signal value of each pixel.

  However, when the image sensor ID is a color image sensor, color filters such as R (red), G (green), and B (blue) are generally arranged in a dot shape. In this case, the spectrum of the light source and each color Due to the spectral transmission characteristics of the filter and the spectral sensitivity characteristics of the image sensor, a difference occurs in the average calibration signal values of the pixels of each RGB color.

  For this reason, in the correction of the color image sensor, the average calibration signal value of each RGB color is obtained from the calibration signal value of each RGB color pixel, and is obtained by dividing this by the calibration signal value of each RGB color pixel. .

  The sensitivity variation for each pixel in the image sensor ID also varies depending on the illumination brightness of the near infrared light source 22, the focal length of the optical system OP in the image camera 23, and the shutter speed of the image sensor ID.

  Therefore, in the case of this embodiment, the memory 24 is set with a plurality of correction patterns each including the illumination brightness of the near-infrared light source 22, the focal length of the optical system OP in the imaging camera 23, and the shutter speed of the imaging element ID. A correction coefficient is stored for each correction pattern.

  In this blood vessel imaging unit 12, when an imaging command is given from the control unit 10, the drive unit 21 irradiates one or more near infrared light sources that irradiate near infrared light to a predetermined position in the blood vessel imaging unit 12. 22 and the image sensor ID in the imaging camera 23 are driven.

  The near-infrared light emitted from the near-infrared light source 22 passes through the finger so as to reflect and scatter the inside of the finger when the finger is placed at a predetermined position of the blood vessel imaging unit 12. Is incident on the image sensor ID via the optical system OP as light for projecting the blood vessel (hereinafter referred to as blood vessel projection light). The imaging element ID photoelectrically converts the blood vessel projection light and outputs the photoelectric conversion result to the drive unit 21 as an image signal at a predetermined period.

  Under the control of the control unit 10, the drive unit 21 adjusts the illumination brightness with respect to the near-infrared light source 22 and adjusts the focal length of the optical system OP and the shutter speed of the imaging element ID in the imaging camera 23. .

  At this time, the drive unit 21 selects from the memory 24 correction coefficients corresponding to the illumination brightness of the near-infrared light source 22 at the present time, the focal length of the optical system OP in the imaging camera 23, and the shutter speed of the imaging element ID. Then, the selected correction coefficient is corrected by multiplying the image signal output from the image sensor ID to cancel the sensitivity variation inherent to the image sensor. The drive unit 21 is configured to send the corrected image signal to the control unit 10.

(2) Specific Processing Contents of Mode Next, the processing contents of the blood vessel registration mode and the authentication mode by the control unit 10 will be specifically described.

(2-1) Blood vessel registration mode When the control unit 10 receives the execution command COM1 via the operation unit 11, the control unit 10 changes the operation mode to the blood vessel registration mode, and controls the blood vessel imaging unit 12 and the blood vessel extraction unit 13.

  Then, the control unit 10 smoothes the image signal sequentially supplied from the blood vessel imaging unit 12 as a blood vessel imaging result of the registrant's finger, and a Gaussian filter or the like in the blood vessel extraction unit 13 for the image signal. Based on each binary image obtained by the binarization process and the thinning process, a process for detecting an optimal imaging condition and extraction condition for the blood vessel pattern (hereinafter referred to as an optimal condition determination process) is executed. .

  Thereafter, the control unit 10 controls the blood vessel imaging unit 12 to perform imaging under the imaging condition determined by the optimum condition determination process, and outputs an image signal output from the imaging element ID as an imaging result under the imaging condition. On the other hand, the blood vessel extraction unit 13 is controlled to extract a blood vessel pattern in accordance with the extraction condition determined by the optimum condition determination process.

Then, the control unit 10 identifies the registrant by associating the imaging conditions and the extraction conditions and the blood vessel pattern extracted under the conditions with the identifier assigned at this time (hereinafter referred to as an assignment identifier). Information is stored in the flash memory 14 as information DRE (hereinafter referred to as registration data) _DRE .

  Further, the control unit 10 is configured to transmit the allocation identifier to the registrant via the notification unit 15. Specifically, the allocation identifier is transmitted to the registrant via a card-like storage medium or display on the display unit.

  In this way, the control unit 10 can execute the blood vessel registration mode.

(2-2) Authentication mode In this authentication device 1, when the person to be authenticated unlocks the door to which the door unlocking device is connected, the person to be authenticated registers with the input means (not shown) in the door unlocking device. Enter the assigned identifier that was sometimes communicated. For example, when assigned identifier vascular imaging unit 3 ₁ as an object is input, the assignment identifier, from the door unlocking device is provided to control unit 10 of the execution instruction COM2 authentication unit 2 via the communication unit 16 as.

In this case, the control unit 10 transitions to the authentication mode, and reads registration data D _RE (imaging condition, extraction condition, and blood vessel pattern) associated with the assigned identifier in the execution command COM2 from the flash memory 14.

The control unit 10, the time readout imaging condition, and sends through the communication unit 16 to the blood vessel imaging unit 3 ₁ mounted on the door unlocking device is a delivery source of the execution command COM2, in the imaging condition Requests to be imaged.

The blood vessel imaging unit 3 _1, the configuration including the same drive unit and the blood vessel imaging section 12, a near infrared light source and the imaging camera, and a memory correction coefficients are stored for each correction pattern in the blood vessel imaging unit 3 ₁ The image signal obtained by imaging the blood vessel in the finger of the person to be authenticated under the imaging condition is multiplied by the correction coefficient corresponding to the requested imaging condition, and the image signal thus obtained is transmitted via the communication unit 16. To the control unit 10.

Incidentally, the blood vessel image pickup unit 3 ₂ to 3 _n, similarly to the blood vessel imaging unit 3 _1, and the correction coefficient in the blood vessel imaging unit 3 is held in each correction pattern, when receiving a request to be imaged The image signal obtained by imaging a blood vessel in the finger of the person to be authenticated under the imaging condition can be multiplied by a correction coefficient corresponding to the imaging condition requested at that time.

Control unit 10 subjects the image signal supplied from the blood vessel imaging unit 3 _1, according to the extraction condition read from the flash memory 14, the blood vessel registration mode same extraction process and through the blood vessel extraction unit 13 performs.

In this state, the control unit 10 collates the extracted blood vessel pattern with the blood vessel pattern of the registration data _DRE read from the flash memory 14, and the user who placed the finger at this time determines whether or not the registrant It is determined whether or not (regular user).

Here, the control unit 10, when it is determined that the registered person, to the door unlocking device for mounting a blood vessel imaging unit 3 _1, and generates an execution command COM3 to the effect that performs the unlocking operation, it The data is transferred via the communication unit 16.

In contrast, the control unit 10, when it is determined that it is not a registrant, to that effect, through the display means of the door unlocking device for mounting a blood vessel imaging unit 3 ₁ adapted to notify the subject of authentication ing.

  In this way, the control unit 10 can execute the authentication mode.

(3) Specific Contents of Optimal Condition Determination Process Next, specific contents of the optimal condition determination process executed by the control unit 10 in the blood vessel registration mode will be described.

(3-1) Determination of Optimal Illuminance Luminance For example, as shown in FIG. 2, the control unit 10 makes the focal length of the optical system OP in the imaging camera 23 with respect to the blood vessel imaging unit 12 when transitioning to the blood vessel registration mode. , and causes set to the representative value the shutter speed of the image pickup device ID, and the illumination brightness for near infrared light sources 22, is set on the illumination luminance LU ₁ the lowest.

The control unit 10, the illumination brightness LU ₁ to image the blood vessels of the registrant finger, the n images IM1 ₁ ~IM1 _n outputted at a predetermined period from the imaging device ID as the imaging result, the time axis To calculate correlation values CR1 _{1 to} CR1 _(n−1) between adjacent images, and calculate the statistic (hereinafter referred to as correlation statistic) SCR1 of these correlation values CR1 _{1 to} CR1 _(n−1). calculate. For the calculation of this statistic, for example, an average, variance, standard deviation or the like is employed.

The control unit 10, these images _IM1 1 _~IM1 2 value obtained as a blood vessel extraction processing result of the blood vessel extraction unit 13 for _n image MIM1 ₁ ~MIM1 blood vessel pattern of _n (the number of pixels constituting the blood vessel pattern) PQ1 ₁ ˜PQ1 _n is calculated, and a statistic (hereinafter referred to as a pattern statistic) SPQ1 of the blood vessel pattern quantities PQ1 _{1 to} PQ1 _n is calculated.

Next, the control unit 10 keeps the focal length of the optical system OP in the imaging camera 23 and the shutter speed of the imaging element ID as representative values, and increases the illumination luminance for the near-infrared light source 22 by a predetermined width. The correlation statistics SCR2 of the correlation values CR2 _{1 to} CR2 _(n−1) between the _n images IM2 _{1 to} IM2 n captured at this time, and the binary images MIM2 _{1 to} MIM2 after the blood vessel extraction processing for the images calculating a blood vessel pattern quantity _PQ2 1 _{~PQ2 n} pattern statistics SPQ2 in _n.

In this way, the control unit 10 sets the illumination brightness with respect to the near-infrared light source 22 by successively increasing the predetermined brightness until the maximum illumination brightness LU _n is reached, and _n captured with the set illumination brightness LU. The correlation statistic SCR of the correlation value CR in one image and the pattern statistic SPQ of the blood vessel pattern amount PQ in the binary image are calculated.

Eventually, the correlation statistics SCRm of the correlation values CRm _{1 to} CRm _{(n−1) in} the _n images IMm _{1 to} IMm _n captured with the highest illumination luminance LU _m and the binary images MIMm _{1 to} MIMm _n When a pattern statistic SPQm of vascular pattern amount _PQm 1 _{~PQm n} finishes calculated, the control unit 10, the correlation statistics SCR1~SCRm and corresponds to the illumination brightness _LU 1 _{~LU m} of each stage was determined heretofore The optimal illumination brightness is determined from the distribution of the pattern statistics SPQ1 to SPQm.

  Here, FIG. 3 shows a distribution example of correlation statistics SCR1 to SCRm (FIG. 3A) and a distribution example of pattern statistics SPQ1 to SPQm (FIG. 3B). As can be seen from FIG. 3, both the correlation statistic and the pattern statistic increase due to the influence of thermal noise as the illumination luminance decreases, and decrease as the illumination luminance increases because the state approaches a saturated state.

  Accordingly, the increase start point P1 where the correlation statistic continues to increase or decrease, the range AR1 between the decrease start point P2, and the increase start point P3 where the pattern statistic continues to increase or decrease, and the decrease start point It can be said that the range AR2 (FIG. 3B) between P4 is stable. (Hereinafter, the range AR1 is called the correlation stable range AR1, and the range AR2 is called the pattern stable range AR2.)

  However, since there are individual differences in the amount of near-infrared light transmitted through the finger and the amount of blood vessels inside the finger, even if the stable correlation range AR1 and the stable pattern range AR2 are stable, they vary. For example, when a finger has a good near-infrared light transmission amount, but the amount of blood vessels inside the finger is small, the pattern stable range is roughly compared to the correlation stable range AR1, as in the example of FIG. AR2 becomes larger. On the other hand, if the finger has a poor near-infrared light transmission amount, but the amount of blood vessels inside the finger is large, the correlation stable range AR1 is larger than the pattern stable range AR2.

  As described above, the magnitude (width) of the correlation stable range AR1 and the pattern stable range AR2 varies due to individual differences. Therefore, if the optimum value is determined by focusing only on the correlation stable range AR1, the optimum Even if the registrant's finger is imaged with the illumination brightness of the value, the amount of blood vessel pattern extracted from the imaging result may be reduced due to, for example, poor transmission of the finger. It lacks reliability as a value.

  On the other hand, when the optimum value is determined by paying attention only to the pattern stable range AR2, the same may be true. In this case, the vicinity of the increase start point P3 is influenced by thermal noise, or the heat Since it is unclear whether the blood vessel pattern amount itself is large rather than the influence of noise, it is not preferable to determine the optimum value.

Therefore, in the case of this embodiment, the control unit 10 determines the correlation stable range AR1 (FIG. 3A) between the increase start point P1 and the decrease start point P2 in which the correlation statistic continues to increase or decrease, and the pattern. The illumination luminance LU _BT corresponding to the intermediate position of the overlapping range AR _RP among the pattern stable ranges AR2 (FIG. 3 (B)) between the starting points P3 and P4 whose statistics are continuously increasing or decreasing, The optimum illumination brightness is determined.

(3-2) Determination of the optimal exposure time Next, the control unit 10 has determined such optimal illumination luminance LU _BT, with respect to blood vessel imaging section 12, the illumination brightness for near infrared light sources 22, the after setting the optimal illumination luminance LU _BT, as shown in FIG. 4, the shutter speed of the imaging device ID (exposure time), sequentially set by a predetermined value pulled from the lowest value EX ₁ up value EX _m.

Further, as in the case of the above-described illumination luminance, the control unit 10 performs correlation statistics of correlation values CRm _{1 to} CRm _(n−1) in _n images IMm _{1 to} IMm _n for each of the stages EX _{1 to} EX _m. and amount SCRM, respectively calculates the pattern statistics SPQm vascular pattern amount _PQm 1 _{~PQm n} in the binary image MIMm ₁ ~MIMm _n obtained as the predetermined blood vessel extraction processing result for the image.

The control unit 10, the shutter speed (exposure _time) of each stage EX 1 ~EX of distribution of the correlation statistic SCR1~SCRm and pattern statistics SPQ1~SPQm corresponding to _m, that ever-increasing or decreasing The shutter speed (exposure time) corresponding to the intermediate position of the overlapping range between the starting points is determined as the optimum shutter speed (exposure time).

(3-3) Determination of Optimal Smoothing Filter Coefficient Next, when the control unit 10 determines the optimal shutter speed (exposure time), the controller 10 determines the shutter speed of the image sensor ID for the blood vessel imaging unit 12. After the optimum shutter speed is set, as shown in FIG. 5, the smoothing filter coefficient in the smoothing process in the blood vessel extraction process is changed from the lowest value FC ₁ (the coefficient value corresponding to the roughest smoothing) to the highest. The value FC _m (coefficient value corresponding to the finest smoothing) is increased by a predetermined width and sequentially set.

The control unit 10 includes, as in the case of illumination brightness and shutter speed above (exposure time), for each of these stages _FC 1 _{~FC m,} correlation values in n images _IMm 1 _{~IMm n CRm} 1 _{~CRm (} calculating a correlation statistic SCRm of _n-1), and a pattern statistics SPQm vascular pattern amount _PQm 1 _{~PQm n} binary image MIMm within _{1 ~MIMm n} obtained as the predetermined blood vessel extraction processing result for the image, respectively To do.

The control unit 10 of the distribution of the correlation statistic SCR1~SCRm and pattern statistics SPQ1~SPQm corresponding to the smoothing filter coefficients _FC 1 _{~FC m} of each stage, so that the ever-increasing or decreasing its A smoothing filter coefficient corresponding to an intermediate position in a range where ranges between start points overlap is determined as an optimum smoothing filter coefficient.

  In this way, the control unit 10 can determine the illumination brightness and the exposure time as the imaging conditions, and can determine the smoothing filter coefficient as the extraction condition.

(4) Operation and Effect In the above configuration, the control unit 10 in the authentication device 1 includes the luminance condition of the near infrared light source 22 that irradiates near infrared light as imaging light, the imaging condition of the imaging camera 23, Among the extraction conditions in the blood vessel extraction unit 13 that extracts the blood vessel pattern from each image signal sequentially output from the image sensor ID as the imaging result of the blood vessel pattern, other than the luminance condition of the near-infrared light source 22 is fixed. The condition (illumination luminance) is switched to a plurality of stages LU _{1 to} LU _m (FIG. 2) via the drive unit 21.

Further, the control unit 10 correlates the correlation statistic SCRm of the correlation values CRm _{1 to} CRm _{(n−1) in} the _n images IMm _{1 to} IMm _n and the binary image MIMm obtained as a result of predetermined blood vessel extraction processing on the image. Request and ₁ pattern statistics of the blood vessel pattern amount _PQm 1 _{~PQm n} in ~MIMm _n SPQm every stage _LU 1 _{~LU m} (Figure 2).

Then, the control unit 10 increases the increase start points P1 and P3 and the decrease start points that follow the increase and decrease in the distribution (FIG. 3) of the correlation statistics SCRm and the pattern statistics SPQm in each stage LU _{1 to} LU _m . The illumination luminance LU _BT corresponding to the intermediate position of the range AR _RP where the ranges AR1 and AR2 between P2 and P4 overlap is determined as one of the conditions suitable for the blood vessel pattern.

  The distribution of the correlation statistic SCRm (FIG. 3 (A)) roughly reflects individual differences and imaging conditions for the individual differences, and is obtained by adding the extraction conditions for the individual differences to the pattern statistics. This is reflected in the distribution of the quantity SPQm (FIG. 3B).

Thus the control unit 10, as compared with the case of determining the illumination brightness LU _BT suitable for blood vessel pattern by focusing only on one of the distribution of the correlation statistic SCRm and pattern statistics SPQm, focusing on both the distribution better to determine the illumination brightness LU _BT suitable vessel pattern Te is more further can determine the illumination brightness LU _BT as suitable for the blood vessel pattern, be thereby obtaining a biological pattern capable of improving authentication accuracy It becomes possible.

In the same manner as the determination of the illumination brightness LU _BT, the control unit 10 determines as one of the conditions suitable for the blood vessel pattern also exposure time of the image pickup device ID, which is one of the components of the imaging camera 23, also Not only the imaging camera 23 (imaging system) but also the filter coefficient in the smoothing process, which is one of the components of the blood vessel extraction unit 13 (image processing system), is determined as one of the conditions suitable for the blood vessel pattern.

  Therefore, even if the authentication apparatus 1 is configured using the imaging camera 23 and the blood vessel extraction unit 13 from different manufacturers, it is possible to obtain a biometric pattern that can further improve the authentication accuracy.

In the case of this embodiment, the control unit 10 is obtained by dividing the average value of each pixel value of an image signal obtained by imaging a predetermined uniform imaging target with the imaging element ID by each pixel value. Each image signal is multiplied by a correction coefficient, a correlation statistic SCRm is obtained based on each multiplied image signal, and a binary image MIMm obtained as a predetermined blood vessel extraction process result for each multiplied image signal obtaining a pattern statistics SPQm from _{1 ~MIMm n.}

  Therefore, it is possible to determine conditions suitable for the blood vessel pattern in consideration of not only individual differences but also sensitivity variations unique to the imaging element ID, and it is possible to obtain a biometric pattern that can further improve the authentication accuracy. .

In this case, even if any blood vessel imaging unit 3 _{1 to} 3 _{N is used} for imaging, imaging can be performed under the same conditions without depending on sensitivity variations unique to the imaging element ID. The authentication apparatus 1 that can further improve the authentication accuracy can be realized.

According to the above configuration, the distribution of the correlation statistic SCRm (FIG. 3 (A)) reflecting the individual difference and the imaging condition for the individual difference, and the extraction condition for the individual difference are added to these. In the reflected distribution of pattern statistics SPQm (FIG. 3B), ranges AR1 and AR2 overlap between the increase start points P1 and P3 and the decrease start points P2 and P4 that continue to increase and decrease. By obtaining the illumination brightness LU _BT , exposure time, and smoothing filter coefficient corresponding to the intermediate position of AR _RP as one of the conditions suitable for the blood vessel pattern, a biological pattern that can improve the authentication accuracy is obtained. Thus, the authentication device 1 that can improve the authentication accuracy can be realized.

(5) Other Embodiments In the above-described embodiments, the case where a blood vessel pattern is applied as a biological pattern has been described. However, the present invention is not limited to this, and for example, a fingerprint pattern, a lip pattern, or A face pattern or the like may be applied.

  In addition, when applying a fingerprint pattern, a lip pattern, a face pattern, or the like, the near-infrared light source 22 that irradiates near-infrared light as imaging light is not necessary, and as in the above-described embodiment, Since it becomes unnecessary to obtain the illumination brightness of the near-infrared light source 22 suitable for the biological pattern, the configuration of the authentication device 1 as a whole can be simplified.

In the above-described embodiment, for example, as an illumination luminance determination method suitable for the blood vessel pattern, among the distribution of the correlation statistic SCRm and the pattern statistic SPQm (FIG. 3) in each stage LU _{1 to} LU _m , the increase and The illumination brightness LU _BT corresponding to the intermediate position of the range AR _RP where the ranges AR1 and AR2 between the increase start points P1 and P3 and the decrease start points P2 and P4 that continue to decrease are overlapped, is the present invention. not limited, in consideration of the predetermined margin with respect to the intermediate position may be a lighting luminance LU _BT corresponding to any position of the midpoint.

Further, in the aforementioned embodiment, with multiple blood vessel imaging unit 3 _N individual door unlocking device has dealt with the case of authentication when unlocking the door, the present invention is to For example, it may be mounted on a portable terminal such as a cellular phone or PDA (Personal Digital Assistants), and authentication may be performed when a restriction on a specific function in the portable terminal is released. Moreover, this optimal condition determination process is applicable not only to this example but to various other authentication modes.

  Further, in the above-described embodiment, the case where the smoothing process, the binarization process, and the thinning process are performed as the extraction unit that extracts the biological pattern from the image signal has been described, but the present invention is not limited thereto. Instead, some of these processes may be omitted or replaced, or new processes may be added to these processes. Incidentally, the order of these processes can be changed as appropriate.

Further, in the above-described embodiment, the case where the flash memory 14 is applied as a storage medium to be registered with the registration data _DRD has been described. However, the present invention is not limited to this, and for example, an HDD (Hard Disk Drive ) And a removable memory stick (registered trademark of Sony Corporation), and various other storage means can be widely applied. In this storage medium, the authentication unit 2 may be separated from the authentication unit 2 and connected via a wireless transmission path.

  Further, in the above-described embodiment, the case where the authentication device 1 having the verification function and the registration function is applied has been described. However, the present invention is not limited to this, and the mode is divided into a single device for each function. It can be applied in various modes depending on the application.

  The present invention can be used in the field of biometric authentication.

It is a block diagram which shows the whole structure of the authentication apparatus by this Embodiment. It is a basic diagram with which it uses for description of the correlation value between images in each illumination brightness | luminance, and the statistics of the blood vessel pattern amount. It is a basic diagram with which it uses for description of determination of optimal illumination luminance. It is a basic diagram with which it uses for description of the correlation value between images in each exposure time, and the statistics of the blood vessel pattern amount. It is a basic diagram with which it uses for description of the correlation value between images in each smoothing filter coefficient, and the statistics of the blood vessel pattern amount.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Authentication apparatus, 10 ... Control part, 11 ... Operation part, 12 ... Blood vessel imaging part, 13 ... Blood vessel extraction part, 14 ... Flash memory, 15 ... Notification part, 16 ... Communication part, 21... Drive unit, 22... Near-infrared light source, 23... Imaging camera, 24... Memory, ID.

Claims (6)

Of the imaging conditions of the imaging unit and the extraction conditions for extracting the biological pattern from each image signal sequentially output as the imaging result of the biological pattern in the imaging unit, except for the exposure time of the imaging element included in the imaging unit A first step of fixing and switching the exposure time to a plurality of stages;
For each of the image signals, a statistic of a correlation value between images adjacent to each other on the time axis is obtained for each stage, and a statistic of a biological pattern amount extracted from the image signal is obtained for each stage. And the steps
In the distribution of the statistics of the correlation value and the statistics of the biometric pattern amount in each of the above stages, the intermediate position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlaps A condition determining method comprising: a third step of determining an exposure time corresponding to a neighboring position as an imaging condition suitable for the biological pattern. Multiplication of multiplying each image signal by a correction coefficient obtained by dividing an average value of each pixel value of an image signal obtained by imaging a predetermined uniform imaging object with the image sensor by the pixel value. With steps,
In the second step,
For each multiplied image signal, a statistical value of a correlation value between images adjacent to each other on the time axis is obtained for each stage, and a statistical amount of a biological pattern extracted from each multiplied image signal is calculated as described above. The condition determination method according to claim 1, wherein the condition determination method is obtained for each stage. In the first step,
Among the luminance conditions of the light source that irradiates near-infrared light as imaging light, the imaging conditions of the imaging means, and the extraction conditions, other than the exposure time of the imaging element or the luminance of the light source is fixed, and a plurality of the exposure times or luminances are set. Switch to each stage,
In the second step,
For each of the image signals, a statistical value of a correlation value between images adjacent to each other on the time axis is determined for each stage, and a statistical amount of a blood vessel pattern extracted from the image signal is determined for each stage.
In the third step,
In the distribution of the statistics of the correlation value and the statistics of the biometric pattern amount in each of the above stages, the intermediate position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlaps The condition determination method according to claim 1, wherein an exposure time and luminance corresponding to a neighboring position are determined as conditions suitable for the biological pattern. In the first step,
Among the imaging conditions and the extraction conditions, the exposure time of the image sensor or a smoothing filter coefficient other than the smoothing filter coefficient of the smoothing process applied to the image signal is fixed, and the exposure time or the smoothing filter coefficient is switched to a plurality of stages, respectively.
In the third step,
In the distribution of the statistics of the correlation value and the statistics of the biometric pattern amount in each of the above stages, the intermediate position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlaps The condition determination method according to claim 1, wherein an exposure time and a smoothing filter coefficient corresponding to a neighboring position are determined as conditions suitable for the biological pattern. An extraction means for extracting the biological pattern from each image signal sequentially output from the image sensor included in the imaging means as an imaging result of the biological pattern;
Control means for controlling the imaging means and the extraction means,
The control means includes
Of the imaging conditions of the imaging means and the extraction conditions of the extracting means, other than the exposure time of the imaging element is fixed, the exposure time is switched to a plurality of stages,
For each image signal, a statistical value of a correlation value between images adjacent to each other on the time axis is obtained for each stage, and a biological pattern quantity statistic extracted from the image signal is obtained for each stage.
In the distribution of the statistics of the correlation value and the statistics of the biometric pattern amount in each of the above stages, the intermediate position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlaps The exposure time corresponding to the vicinity position is determined as an imaging condition suitable for the biological pattern,
One or more biological patterns extracted from each image signal output from the image sensor as an imaging result of the biological pattern under the imaging conditions are registered in a storage medium. Control means for controlling the imaging means and the extraction means for extracting the biological pattern from each image signal sequentially output from the imaging element included in the imaging means as the imaging result of the biological pattern,
Of the imaging conditions of the imaging means and the extraction conditions of the extraction means, a first process that fixes other than the exposure time of the imaging device and switches the exposure time to a plurality of stages,
For each of the image signals, a statistic of a correlation value between images adjacent to each other on the time axis is obtained for each stage, and a statistic of a biological pattern amount extracted from the image signal is obtained for each stage. And processing
In the distribution of the statistics of the correlation value and the statistics of the biometric pattern amount in each of the above stages, the intermediate position of the range where the range between the increase start point and the decrease start point that continues to increase and decrease overlaps A third process for determining an exposure time corresponding to a neighboring position as an imaging condition suitable for the biological pattern;
And a fourth process of registering one or more biological patterns extracted from each of the image signals output from the image sensor as a result of imaging the biological pattern under the imaging conditions. A program characterized by

Images