JPH0235577A - Individual certifying device - Google Patents

Abstract

PURPOSE:To accurately obtain the length information between the joints of the finger of an individual by adding vertical directional density for the longitudinal direction of the picture pattern of the finger and calculating the length information between the joints of the finger using the position of the minimum value of the added value. CONSTITUTION:The input picture of a finger F inputted to a photoelectric converting part 5 is converted into an electric signal by a photoelectric converting element 6 and temporarily stored through an amplifier 7 and an A/D converter 8 in a picture memory 9. In a collation deciding part 11, the length information between the joints of each finger F is obtained from the picture pattern of the finger F stored in the picture memory 9, and it is con firmed that the obtained length information is the length information of the object individual by comparing the obtained length information with the length information between the joints of the finger registered in a feature parameter dictionary part 10. The length information between the joints of the finger F is calculated by adding the vertical directional density for the longitudinal direction of the inputted picture pattern of the finger F and using the position of the minimum value of the added value.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Field of Industrial Application) The present invention relates to a personal authentication device that authenticates an individual using, for example, a bill.

(Prior Art) Recently, an authentication device that uses an individual's handprint has been developed as an authentication device for authenticating an individual.

As this type of personal authentication device, for example,
There is one described in Japanese Patent No. 143222. This personal authentication device reads the handprint image of a hand placed on a hand rest, converts it into an electrical signal, digitizes this electrical signal, and extracts characteristic parameters from the obtained handprint image pattern. By comparing the image pattern with the characteristic parameters extracted from the image pattern of the individual's handprint, a matching judgment, that is, identity verification is performed.

By the way, the personal authentication device described in J1 binarizes the image pattern of the obtained handprint using an appropriate threshold value (S1notschhold level), and binarizes it in the longitudinal direction of the finger corresponding to the joint of the finger. The length information between finger joints obtained by detecting vertical wrinkles is used as a feature parameter.

However, in the method described above that obtains length information between finger joints, it is difficult to set the threshold for binarization, and if the threshold is set incorrectly, wrinkles corresponding to the finger joints may not be detected. In some cases, extra wrinkles other than the wrinkles corresponding to the finger joints were detected. For this reason, the detection rate of length information between finger joints is unstable, resulting in low accuracy of personal authentication.

(Problem to be Solved by the Invention) This invention is a method for obtaining length information between finger joints by binarizing an image pattern of a handprint and detecting wrinkles corresponding to the finger joints. It was difficult to set the threshold value for This was created to solve the problem that the detection rate of length information was unstable and the accuracy of personal authentication was low. Although it has a relatively simple configuration, it does not rely on human-powered images. We have developed a personal authentication device that can stably detect wrinkles corresponding to finger joints, thereby making it possible to accurately and reliably obtain length information between finger joints and improving authentication accuracy. The purpose is to provide.

[Structure of the invention]

(Means for Solving the Problems) The personal authentication device of the present invention includes a photoelectric conversion device that reads an input image of a finger and converts it into a corresponding electrical signal, and an image of the finger that is an output of the photoelectric conversion device. storage means for storing patterns; storage means for storing characteristic parameters extracted from pre-registered finger image patterns of individuals; and storage means for storing characteristic parameters extracted from finger image patterns stored in the storage means; a personal authentication device that uses length information between the joints of a finger as the feature parameter, the personal authentication device includes a collation unit that extracts the information and matches it with a feature parameter stored in the storage unit, and uses length information between finger joints as the feature parameter; The structure includes calculation means for calculating length information between the joints of the fingers using the position of the minimum value of the integral signal obtained by adding the density in the direction perpendicular to the longitudinal direction of the pattern. .

(Operation) This invention provides a personal authentication device that uses length information between finger joints as a feature parameter, in which a human input image of a finger is read and converted into an electrical signal in a direction perpendicular to the longitudinal direction of the image pattern of the finger. An integral signal is obtained by adding the concentrations of This allows information to be entered accurately and reliably.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 5 shows the personal authentication device of the present invention.

In FIG. 5, reference numeral 1 denotes a finger rest on which a finger F is placed, for example, for inputting an image pattern of a finger (for example, the four finger parts of the index finger, middle finger, ring finger, and little finger), and is, for example, a transparent acrylic plate or a transparent glass plate. It is made of transparent materials such as The input image of the finger F placed on the finger rest 1 is generated by the irradiation of light from the light FA3 disposed below the finger rest 1 through the finger rest 1. The reflected light image is formed as a reflected light image, and is input to a photoelectric conversion section (photoelectric conversion means) 5 via an optical system 4. The input image of the light finger input to the photoelectric conversion unit 5 is, for example, CC
It is converted into a corresponding electrical signal by a photoelectric conversion element 6 such as a D-shaped area sensor or line sensor. This electrical signal is amplified by an amplifier 7 and then A/D converted by an A/D converter 8 to be digitized. This digitized finger image pattern is temporarily stored in an image memory (storage means) 9.

On the other hand, the feature parameter dictionary section (storage means) 10 stores feature parameters extracted from pre-registered individual finger image patterns. This feature parameter is length information between the joints of the fingers, for example, the distance between the first joint and the second joint of the index finger, middle finger, ring finger, and little finger, or (and/or) the distance between the second joint and the third joint. is used.

The comparison determination unit (verification means) 11 obtains length information P between the joints of six fingers from the image pattern of the fingers stored in the image memory 9.
]22 The identity is verified by comparing with the length information Q between the joints of the fingers registered in the feature parameter dictionary section 10.

The host machine 12 is configured to control, for example, opening/closing a door, sounding an alarm buzzer, etc. in response to a signal from the verification/judgment section 11.

Next, the flow of processing in the collation determination section 11 will be explained with reference to the flowchart shown in FIG.

First, an area corresponding to six fingers is cut out from the finger image pattern stored in the image memory 9.

This process, as shown in FIG. (indicated by a chain line) is extracted.

When a pattern as shown in FIG. 3 is extracted by cutting out the area of the six fingers, detection in the longitudinal direction of the six fingers is performed. This is obtained by approximating the outer circumference of the side surface of the finger to a straight line S using the method of least squares or the like.

When the longitudinal direction of the finger is detected in this way, the distance between each joint of the cut out finger is calculated.

In this case, as shown in FIG. 4, an integral signal 41 is obtained by adding the image density in a direction perpendicular to the longitudinal direction of the obtained finger image f. Since this integral signal 41 takes a minimum value at the position of the horizontal crease corresponding to the joint of the finger, the distance between each joint can be calculated by finding the distance between these minimum values.

That is, the point where the integral signal 41 exceeds a certain threshold TH for the first time is defined as the finger tip position Xe, and a search range L1 for finding the minimum value of the integral signal 41 is set with this tip position 1fXe as a reference. The setting of this search range L1 utilizes the fact that the length between the joints of human fingers is determined to some extent. Then, the minimum value of the integral signal 41 within the search range L1 is searched, and the position of this minimum value is set as Xl.

Next, with this minimum value position X1 as a reference, a search range L2 is set using the fact that the length between the joints of human fingers is determined to some extent. Then, the minimum value of the integral signal 41 within this search range L2 is searched, and the position of this minimum value is set as X2.

The distance Il between the position X1 and the position x2 of the minimum value found in this way
By determining l (X2-X+), the distance (P) between the first joint and the second joint is calculated. Further, the distance between the second joint and the third joint can also be calculated in the same manner.

The feature parameter P obtained in this way is not susceptible to variations in the quality of the input image, such as how the fingers are placed, and is also used as a threshold for binarization to detect horizontal wrinkles corresponding to the joints of the fingers. It is considered a stable parameter because it does not require any setting.

After extracting the feature parameters P for the six fingers from the image pattern in the image memory 9 in this way, the feature parameters Q registered in the feature parameter dictionary section 10 are compared with the extracted feature parameters P. , a comparison determination is made. In this case, the feature parameters extracted from the input finger image pattern are set to P1, the feature parameters registered in the feature parameter dictionary section 10 are set to Q, and, for example, a judgment value that is an experimentally determined criterion is set to R. Then, the difference P- between the parameters P and Q above is
The determination is made based on the magnitude relationship between Q and the determination value R. In other words, R≦I P-Q I Match R> I P-Q I Mismatch As a result of this comparison judgment, if it is determined that there is a match, a match signal will be sent to the host, and if it is determined that there is no match, a mismatch signal will be sent to the host. Output to machine 12.

Then, the host machine 12 performs processing according to the signals, such as opening a door in response to a matching signal, and sounding an alarm buzzer in response to a non-coinciding signal.

As mentioned above, in a personal authentication device that uses length information between finger joints as a feature parameter, the density in the direction perpendicular to the longitudinal direction of the finger image pattern that is obtained by reading the input finger image and converting it into an electrical signal. By adding up the integral signal and calculating the distance between the joints of the fingers using the position of the minimum value of this integral signal, the distance between the joints of the fingers as an individual characteristic parameter can be accurately and Make sure you can enter it correctly.

That is, by adding the density of the finger image in a direction perpendicular to the longitudinal direction of the finger to obtain an integral signal, horizontal wrinkles corresponding to the finger joints are emphasized, and these horizontal wrinkles are added to the integral signal. Stable detection is achieved by detecting the position of the minimum value. As a result, the length information between the joints of (h) is
It becomes possible to obtain accurately and reliably. Therefore, since stable verification processing can be performed reliably, the verification rate increases and authentication accuracy can be improved.

Note that the accuracy of authentication is proportional to the number of feature parameters to be compared and verified. For this reason, in the above embodiment, the explanation is given using the four fingers of the index finger, middle finger, S finger, and little finger as an example. You will not receive any.

It goes without saying that various other modifications can be made without departing from the gist of the invention.

[Effects of the Invention] As detailed above, according to the present invention, wrinkles corresponding to finger joints can be stably detected regardless of the quality of the input image, although the configuration is relatively simple. Therefore, it is possible to accurately and reliably obtain length information between finger joints, and it is possible to provide a personal authentication device that can improve authentication accuracy.

[Brief explanation of the drawing]

The drawings show one embodiment of the present invention, and FIG. 1 is a flowchart for explaining the processing of the matching determination section, and FIG. 2 is for explaining the processing for cutting out a finger region from a finger image pattern. Figure 3 is a diagram for explaining the process of detecting the longitudinal direction of the finger, Figure 4 is a diagram for explaining the process of calculating the distance between the joints of the fingers, and Figure 5 is a diagram for explaining the process of calculating the distance between the joints of the fingers. It is an overall configuration diagram. 5... Photoelectric conversion section (photoelectric conversion means), 9... Image memory (storage means), 10... Feature parameter dictionary section (
storage means), 11... verification determination section (verification means), 41
...Line segment signal. Applicant's agent Patent attorney Shikihiko Suzue Figure 1 P2 Figure II 3 Figure 5

Claims (1)

[Scope of Claims] Photoelectric conversion means for reading an input image of a finger and converting it into a corresponding electrical signal; storage means for storing an image pattern of the finger which is an output of the photoelectric conversion means; and a pre-registered individual. storage means for storing characteristic parameters extracted from the finger image pattern of the person; a personal authentication device that uses length information between finger joints as the characteristic parameter, the personal authentication device comprising: a verification means for verifying the input finger image pattern; 1. A personal authentication device comprising a calculation means for calculating length information between finger joints using the position of the minimum value of the integral signal obtained by performing the following steps.