JPH0371375A - Fingerprint collating device - Google Patents

Abstract

PURPOSE:To quickly perform the collation of fingerprints by carrying out successively the collation of blocks corresponding to each other between the collected fingerprint data and the registered ones and proceeding to the next collation of blocks for only the registered fingerprint data having the coincidence higher than a prescribed degree. CONSTITUTION:A fingerprint reading means 2 reads the fingerprints and outputs the picture data to a fingerprint data arithmetic means 4. The means 4 extracts the features of fingerprints out of the picture data on the fingerprints to compute the fingerprint data and also to divide the fingerprint data into the prescribed number (n) of blocks based on the picture data. These divided blocks are collated with each other via a collation means 8. In this case, a block 1 of the collected fingerprint data TD is collated with each block 1 of registered fingerprint data RD1 - RDm. Then the collation is carried out in the same way between the blocks 2. In this case, only the data on the blocks 1 having the coincidence higher than a prescribed degree are collated with each other for the data RD1 - RDm. As a result, the registered fingerprint data to be collated are reduced as the collating process proceeds to the block (n) from the block 1.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a fingerprint verification device, and particularly to improving the efficiency of verification thereof.

[Prior Art] A fingerprint verification device verifies the identity of a person by verifying fingerprints. Conventional fingerprint verification devices include the following.

Features are extracted from fingerprint image data, and data related to the features (referred to as fingerprint data) are registered in advance for each hole. During fingerprint verification, fingerprint data is collected again. The degree of match with the collected fingerprint data is examined for all registered fingerprint data, and the one with the highest degree of match is selected. In this way, the identity of the person is confirmed.

There is also a method of registering fingerprint data by adding a PIN number at the time of registration. In this case, at the time of fingerprint verification, fingerprint data is collected and a password is input. Then, the corresponding registered fingerprint data is retrieved, and it is determined whether the collected fingerprint data matches this registered fingerprint data. In this way, the identity of the person is confirmed.

[Problems to be Solved by the Invention] However, conventional fingerprint verification devices have the following problems.

In a device that compares all registered fingerprint data with collected fingerprint data, there is a problem in that the comparison takes time. This is especially true when the feature extraction is made more detailed in order to improve matching accuracy.

On the other hand, such a problem does not occur with devices that require the user to enter a password. However, if the user forgets the password, a problem arises in that the verification operation cannot be performed.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems and provide a fingerprint verification device that can quickly perform verification without using a personal identification number.

[Means for Solving the Problems] FIG. 1 shows a block diagram of the overall configuration of a fingerprint matching device according to the present invention. The fingerprint verification device according to the present invention includes a fingerprint reading unit 2 that reads a fingerprint as image information, and a fingerprint data calculation unit that calculates fingerprint data by dividing it into blocks and extracting their features based on the read image information of the fingerprint. 4. A registered fingerprint data storage means 6 which stores the fingerprint data of the person to be verified by dividing it into blocks as registered fingerprint data in advance; A matching means 8 is provided for matching the next block only with respect to the registered fingerprint data for which a match of a predetermined value or more is found.

[Operation] In the above configuration, the fingerprint reading means 2 reads a fingerprint and outputs image data to the fingerprint data calculation means 4. The fingerprint data calculation means 4 extracts features from the fingerprint image data to calculate fingerprint data, and also divides the fingerprint data into predetermined blocks n based on the image data. Therefore, the fingerprint data calculation means 4 outputs collected fingerprint data TD divided into n pieces. The collected fingerprint data TD is compared with the fingerprint data RD, -RD divided into blocks n, stored in the registered fingerprint data storage means 6, by the matching means 8. This comparison is performed, for example, as follows: First, block 1 of the collected fingerprint data TD is compared with block 1 of the registered fingerprint data RD, ~RD. Next, the same comparison is performed for block 2. At this time, among the registered fingerprint data RD, to RD, only those whose data in block 1 matches a predetermined value or more are compared.Hereafter, the matching is performed in the same manner up to block n, Search for the corresponding registered fingerprint data.

As described above, as the verification progresses from block query to n, the amount of registered fingerprint data to be verified decreases, and the verification time can be shortened.

[Embodiment] FIG. 2 shows the hardware configuration of a fingerprint matching device according to an embodiment of the present invention. In this example, the CPU
The function shown in FIG. 1 is realized using l0. The pass line 30 includes a CPU 10, a memory 12, and a fingerprint reader 14.
, control signal exit section 16, display device 20, transmission control section 22
, an external storage device 24 is connected.

The fingerprint reading device 14 collects image data of a placed fingerprint. The image data may be either image data or vector data. The control signal sending unit 16 is
Based on the verification results, the drive system 1 of the door lock release mechanism etc.
8, it provides a drive signal. The display device 20 is
It is a CRT, etc., and displays guidance etc. during use. The transmission control unit 22 is used to transmit data when performing a verification operation on the host computer. The external storage device 24 is a floppy disk drive, a hard disk drive, or the like, and is used to store registered fingerprint data. The CPU 10 controls each section according to a program stored in the memory 12 and performs a collation operation.

This device has two modes: fingerprint data registration and verification. First, the fingerprint data registration mode will be explained.

1. Registration mode - the registration mode program stored in the memory 12,
A flowchart is shown in FIG. First, in step S1, the CPU 1O controls the fingerprint reading device 14 to capture fingerprint image data. The captured image data is stored in the memory 12. Here, it is assumed that image data as shown in FIG. 4 has been imported.

CPU10 determines the center point O of the image and its coordinates based on this image data (step S2).

In order to facilitate processing, the coordinates of the center point O are set as the origin.

Next, CPU10 uses the calculated center point O as a reference,
Determine the X-axis and Y-axis. According to the X and Y axes, it is divided into blocks 1 to 4 as shown in Figure 4.
Step S1).

Next, in step S4, features of the fingerprint image are extracted for each block and fingerprint data is calculated.

In this example, the branch points, end points,
Features are extracted based on the number of circles, etc. (called feature points) and their coordinates.

Here, the fingerprint data of block 1 is shown in FIG. The first data n indicates the number of feature points existing in this block (here, 4). The following data shows the coordinates of the first feature point A1.

The next data indicates the type of feature point A1 (branch, end point, circle, etc.). Below, the coordinates of the feature point A2, the type of the feature point A2, the coordinates of the feature point A3, the type of the feature point A3, the coordinates of the feature point A4, and the type of the feature point A4. This is the fingerprint data of block 1. Block 2
The same applies to 4. Therefore, the overall structure of the fingerprint data is as shown in FIG. Note that the control code at the beginning is for identifying and searching the registrant.

After calculating the fingerprint data as described above, it is stored (step S1). Although storage may be performed in the memory 12, it is preferable to store it in the external storage device 24 in consideration of capacity and safety.

The above-mentioned fingerprint data is registered for all m persons to be verified. Therefore, the fingerprint data will be stored in a state as shown in FIG. The first line is the registered fingerprint data RD of the first person, the second line is the registered fingerprint data RD of the second person, and hereafter, the registered fingerprint data RD of the mth person in the m line are stored. There is. Each person's registered fingerprint data is divided into n (here, 4) blocks. In the figure, each block data is R
It is displayed as D-.

Registration is completed as described above.

-Verification mode- Next, FIG. 8 shows a flowchart of the verification mode.

When a finger is placed on the fingerprint reader 14 for verification, the CP
Ul0 controls the fingerprint reading device 14 to collect fingerprint image data (step S1°).

Thereafter, fingerprint data TD, -TD for each block is calculated in the same manner as steps 32 to S4 in the registration mode (step Sll-513).

Next, in step S14.5111, i and j
Set the initial value of . Here, i is a variable corresponding to the number m of registrants, and j is a variable corresponding to the number n of divided blocks. In any case, assuming i=1 and j=1,
Verification begins.

In step S1g, CPU10 collects the collected data T.
D, and registered data RD, , are matched.

That is, the collected fingerprint data regarding block 1 is matched with the registered fingerprint data regarding block 1 of the first registrant. Matching is performed based on whether feature points of the same type are located at the same coordinates (some errors are allowed). Specifically, the match score MS expressed by the following formula
is calculated and the degree of matching is quantified.

MS=C/T Here, T is the number of registered feature points, and C is the number of matched feature points. Therefore, the match score M
S takes a value between O and 1, and the larger the value, the higher the degree of matching.

Next, it is determined whether the value of the found match score MS exceeds a predetermined threshold (step SF).
. If it exceeds (that is, if the degree of coincidence is high), the current value of i is stored (step S1). If you haven't crossed it, you won't remember it.

Next, in step S0S, it is determined whether the registered data has been compared with all registered data. Since everyone has not finished yet, step S0S is executed again. step s
In is, 1 is added to the value of i (i=2).

Therefore, in step 8□6 and subsequent steps, the second person's registered data RD2. is the subject of verification. At this time, j
(=1) has not changed, so block 1 remains the target.

The above operation is repeated by sequentially increasing i. As a result, only registrants whose match score exceeds a predetermined match score for block 1 are stored. In other words, registrants are narrowed down.

When i equals m (that is, when matching is completed to the last registrant), step S0. Judgment is YES
Then, the process proceeds to step S2°. In step S0O,
It is determined whether all blocks of the divided blocks have been verified. At this point, only the niblock has been completed, so the process advances to step S22. Step S2! In , j is increased by 1 (j=2). That is, the next block, block 2, is targeted. Then, in step 323, only for the stored i (
That is, for block 1 (only those registrants who exceeded a predetermined match score), for block 2, the above steps S16 to Sta are repeated. This further narrows down the registrants.

When step srs is completed, the process proceeds to step S22 via step S2, where j is further increased by 1 (j=3).
, execute step S23 regarding block 3. Here too, step S01 is performed only for the further narrowed down registrants. Therefore, as the number of blocks j increases, the data to be matched decreases, and the matching time is shortened.

The above operation is repeated, and if j=4 (that is, if matching is completed for all blocks), the process advances to step S21 and a final determination is made.

In the final judgment, the one with the highest match score is selected from among the narrowed down registrants. In addition, as a result of narrowing down,
If there are no registrants left, it is determined that there are no applicable registrants. Depending on the presence or absence of the person, the CPU 10 controls the drive system 18 (for example, unlocks the door) via the control signal sending unit 16.

Fingerprint verification is performed as described above.

-Other Embodiments- Note that in the above embodiment, the number of block divisions is 4, but it can be any number.

Further, in the above embodiment, the image data is divided into blocks and then the fingerprint data is calculated, but the fingerprint data may be calculated and then divided into blocks.

Further, in the above embodiment, the fingerprint features are extracted based on the type and coordinates of the feature points, but other features may be extracted.

Further, in the above embodiment, the functions shown in FIG. 1 are realized by software using CPtJlo, but part or all of them may be configured only by hardware.

[Effects of the Invention] The fingerprint verification device according to the present invention calculates fingerprint data divided into blocks according to image information of the fingerprint, sequentially collates corresponding blocks of collected fingerprint data and registered fingerprint data, and performs a predetermined process. It is characterized in that the next block is compared only for the registered fingerprint data for which a match greater than or equal to the value is found. Therefore, it is possible to provide a fingerprint verification device that can quickly perform verification without using a personal identification number or the like.

[Brief explanation of drawings]

FIG. 1 is an overall configuration diagram of a fingerprint verification device according to the present invention, FIG. 2 is a diagram showing a hardware configuration of a fingerprint verification device according to an embodiment of the invention, FIG. 3 is a diagram showing a flowchart of a registration mode, and FIG. Figure 4 shows an example of fingerprint image data, Figure 5 shows one block of fingerprint data, Figure 6 shows the entire fingerprint data, and Figure 7 shows multiple registered fingerprint data. FIG. 8 is a flowchart of the verification mode. Fingerprint reading means Fingerprint data calculation means Registration fingerprint data storage means Verification means

Claims (1)

[Claims] (1) Fingerprint reading means for reading a fingerprint as image information; fingerprint data calculation means for calculating fingerprint data with its features extracted based on the read image information of the fingerprint; fingerprint data of a person to be matched as pre-registered fingerprint data; In the fingerprint verification device, the fingerprint verification device includes a stored registered fingerprint data storage means, a verification means for sequentially verifying the collected fingerprint data and the registered fingerprint data, and the fingerprint data calculation means divides the fingerprint data into blocks according to image information of the fingerprint. The registered fingerprint data storage means divides the registered fingerprint data into blocks and stores the registered fingerprint data, and the collation means sequentially collates the corresponding blocks of the collected fingerprint data and the registered fingerprint data to obtain a predetermined value. A fingerprint verification device characterized in that the next block is verified only for the registered fingerprint data for which a match has been found.