JP2004287563A - Device and method for inputting and collating biological pattern information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a small and inexpensive device for inputting and collating biological pattern information. <P>SOLUTION: Upon the input of a biological pattern to an input part 1, a data processing unit 3 stores an image of the living body pattern for one input into an input image storage part 21, then compares the image of the living body pattern stored in the input image storage part to reconfigured image information stored in a reconfigured image storage part 22, then calculates position coordinates on the reconfigured image storage part which the image of the pattern in the input image storage part should assume, and stores the image of the pattern for the one input into the reconfigured image storage part according to the calculated coordinates. The data processor repeats the above operations a plurality of times while moving the subject of input. Where the reconfigured image has reached a predetermined size, characteristic data are extracted by analysis of the reconfigured image and stored in a characteristic data storage part 23. The reconfigured image for which the process of extracting characteristic data has been over is disposed of. Using the stored characteristic data, a collating/authenticating device 4 determines if the pattern matches the registered living body pattern information. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an input / collation apparatus for biometric pattern information such as fingerprint information suitable for being mounted on a portable terminal or the like.
[0002]
[Prior art]
Generally, a living body has a biological pattern peculiar to the solid such as a fingerprint, and is used as a means for specifying the solid. For example, when an individual is specified using fingerprint data, the following matching method is available.
“One-to-one collation”: a method of identifying a fingerprint by “information such as ID” from a large number of registered fingerprints, collating the fingerprint with an input fingerprint, and identifying an individual.
"1: N collation": a method of collating with a fingerprint input only with fingerprint data from all registered fingerprints and searching for a matching fingerprint.
[0003]
A high degree of matching accuracy is required to realize "one-to-N matching" for finding a matching fingerprint only with the input fingerprint data. That is, the false acceptance rate (the rate at which unregistered fingerprints are erroneously recognized as registered fingerprints) is reduced, and the false rejection rate (registered fingerprints are erroneously recognized as unregistered fingerprints). Must be reduced. For this reason, by using the information on the feature points and the relations in the fingerprint collation, the false acceptance rate and the false rejection rate are reduced, and both security and convenience are provided (for example, see Patent Documents 1 and 2). .
[0004]
[Patent Document 1]
Japanese Patent No. 1289457 (Japanese Patent Publication No. 60-12677) [Patent Document 2]
Japanese Patent No. 1468842 (Japanese Patent Publication No. 63-13226) [Patent Document 3]
JP-A-2002-505778 [0005]
The feature points and the relation will be described with reference to FIG. As shown in FIG. 6A, a swelling portion forming a fingerprint pattern is called a "ridge". The ridge has a cut portion and a branch portion. The cut portion is called an "end point", and the branch point is called a "branch point". The “end point” and the “branch point” are collectively called a “feature point”. In a normal fingerprint, the feature point at the center of the fingerprint is about 50. "Position" and "direction" obtained from this feature point are handled as basic information of the feature point.
[0006]
However, there is a possibility that coincidence will occur by information only on the position and direction of the feature point. Therefore, information called a relation is added. As shown in FIG. 6B, the “relation” is information on the number of “ridges” crossing between a feature point and another feature point. Compared with the case where only the information of the position and the direction is used, the fingerprint collation accuracy is remarkably improved.
[0007]
When this fingerprint input / collation device is mounted on a portable terminal or the like, it is desirable that the fingerprint input unit be as small as possible. However, a fingerprint input unit that is smaller than the size of the finger cannot acquire the fingerprint of the entire finger at once. Therefore, it is necessary to reconstruct partial fingerprint information obtained by moving a finger over an input portion into overall fingerprint information.
[0008]
With reference to FIG. 7, the operation of inputting a fingerprint, reconstructing fingerprint information, and verifying / authenticating according to the related art will be described.
[0009]
When a fingerprint is input to the input unit, one input fingerprint image is stored in the input image storage unit (step S11), and the position coordinates of the input image storage unit are calculated (step S12). Then, the fingerprint image in the input image storage is stored in the reconstructed image storage (step S13). The above steps are repeated until the input is completed (step S14).
[0010]
When the input is completed and the entire fingerprint image is stored in the reconstructed image storage unit, feature data such as feature points and relations is extracted from the reconstructed image. The extracted feature data is stored in the feature data storage unit (Step S15). After the extraction of the feature data is completed, the reconstructed image in the reconstructed image storage unit is discarded (step S16). Thereafter, the feature data stored in the feature data storage unit is passed to the matching / authentication device to perform a matching / authentication process (step S17).
[0011]
[Problems to be solved by the invention]
However, in the conventional technique, the whole image of the fingerprint is stored in the reconstructed image storage unit, and a capacity enough to store the image of the whole fingerprint is required. Increasing the capacity of the reconstructed image storage unit is not preferable because it increases the size, weight, and power consumption of the apparatus.
[0012]
Further, as an example of a small and inexpensive fingerprint input device, a method of measuring a fingerprint structure using a one-dimensional sensor array is disclosed (for example, see Patent Document 3).
[0013]
Accordingly, an object of the present invention is to provide a smaller and less expensive apparatus and method for inputting and matching biometric pattern information while having the same functions as those of the related art.
[0014]
[Means for Solving the Problems]
In order to solve the above-described problem, the present invention is configured such that an input unit having an area smaller than an input target, an input image storage unit that stores image information obtained from the input unit, and a plurality of input images. A reconstructed image storage unit that stores the obtained image information, a feature data storage unit that stores feature data obtained by analyzing the reconstructed image, and when a biological pattern is input to the input unit, a single input is performed. Storing a biological pattern image in the input image storage unit; comparing the biological pattern image stored in the input image storage unit with reconstructed image information stored in the reconstructed image storage unit; Calculating the position coordinates of the biological pattern image to be taken on the reconstructed image storage unit and, based on the calculated coordinates, storing the biometric pattern image for one input in the reconstructed image storage unit Enter When the reconstructed image reaches a predetermined size while repeating the elephant multiple times while being shifted, data processing for analyzing the reconstructed image to extract feature data and storing the feature data in the feature data storage unit Means and a matching / authentication unit for determining whether or not the stored feature data matches the registered biometric pattern information.
[0015]
Furthermore, the data processing means is characterized by discarding the reconstructed image for which the characteristic data extraction processing has been completed.
[0016]
With the above configuration, the storage capacity of the image storage unit can be reduced.
[0017]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, embodiments of the present invention will be described with reference to the drawings.
[0018]
FIG. 1 shows a configuration of a biometric pattern information input / collation device according to the present invention. Reference numeral 1 denotes an input unit for reading a biological pattern such as a fingerprint. Reference numeral 2 denotes a storage device, reference numeral 3 denotes a data processing device, and reference numeral 4 denotes a verification / authentication device for determining whether or not the obtained characteristic data of the biological pattern matches the registered biological pattern information. The area of the biometric pattern reading portion of the input unit 1 is smaller than the input object (a human finger in the case of fingerprint input). The present invention is applicable not only to fingerprints but also to other biometric patterns such as palm prints. Here, a case where the present invention is applied to fingerprint input will be described as an example.
[0019]
FIG. 2 shows an example of the size of the fingerprint input unit according to the first embodiment. In the input unit 1 in FIG. 2, detection elements are arranged with sufficient fineness to acquire a fingerprint as an image, but the total area of the detection part is configured to be smaller than the area of the finger 5 which is an input target. ing.
[0020]
As a method of the detection element, one that optically acquires an image, one that detects a difference in capacitance generated between a ridgeline and a detection element, and a valley and a detection element, one that similarly detects a pressure difference, Although various methods are conceivable, any method can be used as long as a biological pattern can be input, and another method may be used.
[0021]
The storage device 2 includes an input image storage unit 21, a reconstructed image storage unit 22, and a feature data storage unit 23. The input image storage unit 21 stores image information obtained from the input unit 1. The reconstructed image storage unit 22 stores image information reconstructed from a plurality of input images. The feature data storage unit 23 stores feature data obtained by analyzing the reconstructed image.
[0022]
The data processing device 3 includes a coordinate calculation unit 31, a reconfiguration unit 32, and a feature data extraction unit 33. The coordinate calculation means 31 calculates the position coordinates that the image of the input image storage unit 21 should take on the reconstructed image storage unit 22. The reconstruction unit 32 stores the image information of the input image storage unit 21 in the reconstructed image storage unit 22 using the coordinate position calculated by the coordinate calculation unit 31. The characteristic data extracting unit 33 analyzes the image stored in the reconstructed image storage unit 22 to extract characteristic data, and stores the extracted characteristic data in the characteristic data storage unit 23.
[0023]
Next, the operation of the present embodiment will be described with reference to FIG. When a fingerprint is input to the input unit 1, a fingerprint image for one input is stored in the input image storage unit 21 (step S1).
[0024]
FIG. 4 shows an example of a fingerprint image obtained by one input. The solid line 51 represents the ridge line of the fingerprint. Since the input unit 1 has a smaller area than the finger 5 as shown in FIG. 2, the range of the image that can be acquired by one input is only a part of the whole. In FIG. 4, this range is indicated by squares 1a, 1b, and 1c. The shift of each position indicates that the relative position of the finger 5 and the input unit 1 shifts for each input.
[0025]
Next, the coordinate calculation unit 31 compares the fingerprint image stored in the input image storage unit 21 with the reconstructed image information stored in the reconstructed image storage unit 22, and obtains the fingerprint image in the input image storage unit 21. The position coordinates on the power reconstructed image storage unit 22 are calculated (step S2).
[0026]
The calculation of the position coordinates is performed by checking the degree of coincidence of the images. As an example, a case where the position coordinates of 1b in FIG. 4 are calculated will be described. It is assumed that 1a has been input in the past and has already been stored in the reconstructed image storage unit 22. Here, an overlapping region 1ab exists between 1a and 1b. That is, the fingerprint image of the area 1ab in 1b matches the fingerprint image of the area 1ab stored in the reconstructed image storage unit 22. Thus, the position coordinates of 1b on the reconstructed image storage unit 22 can be calculated.
[0027]
When the coordinates are calculated by the coordinates calculating means 31, the reconstructing means 32 stores one input fingerprint image in the reconstructed image storage unit 22 based on the coordinates (step S3).
[0028]
Here, when the reconstructed image has reached a predetermined size (step S4), the reconstructed image is analyzed to extract feature data, and stored in the feature data storage unit 23 (step S5). The predetermined size is determined in advance so as to be smaller than the entire fingerprint image.
[0029]
If the reconstructed image has not reached the predetermined size (step S4), it is necessary to acquire a fingerprint image, so the image is acquired again (step S1), and the above processing is repeated.
[0030]
The reconstructed image for which the characteristic data extraction processing has been completed is unnecessary and is discarded (step S6). That is, the reconstructed image storage unit 22 only needs to have a capacity of a predetermined size (smaller than the entire fingerprint). There is no need for the capacity to store an image of the entire fingerprint.
[0031]
Next, it is determined whether or not a fingerprint has been input to the input unit 1 (step S7). If it has been input, the process returns to the acquisition of the image (step S1), and the above is repeated. By repeatedly performing image acquisition and feature data extraction, it is possible to obtain feature data necessary for performing highly reliable matching.
[0032]
If the input has not been made, it is considered that the user has finished inputting the fingerprint. Therefore, the feature data stored in the feature data storage unit 23 is passed to the collation / authentication device 4, and the collation / registration with the registered fingerprint information is performed. An authentication process is performed (step S8).
[0033]
FIG. 5 shows the size of the fingerprint input unit according to the second embodiment. In the first embodiment, the input unit 1 has an elongated shape as shown in FIG. 2, but the input unit 1 having a non-elongated shape in the present embodiment shown in FIG. 5 may be used. However, it is configured to be smaller than the area of the finger 5. Other configurations and operations are the same as those of the first embodiment.
[0034]
As described above, the fingerprint is described as an example of the biometric pattern. However, in the case of a palm print or the like, a biometric pattern having a larger area must be input, and a palm print information input / collation device is mounted on a small device such as a portable terminal. In such a case, a greater effect can be obtained by applying the present invention.
[0035]
【The invention's effect】
As described above, according to the present invention, it is not necessary to store the entire image of the fingerprint in the reconstructed image storage unit, which has conventionally required the storage of the entire image of the fingerprint, and the capacity of the reconstructed image storage unit is reduced. Since it can be reduced as compared with the technique of the above, the storage device can be reduced in size and simplified. Therefore, it is possible to manufacture and supply a biometric pattern information input / collation device that is smaller and cheaper while having the same functions as the conventional one.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a biometric pattern information input / collation device according to the present invention.
FIG. 2 is a diagram illustrating the size of a fingerprint input unit according to the first embodiment.
FIG. 3 is a flowchart illustrating the operation of the present invention.
FIG. 4 is a diagram showing an example of a fingerprint image obtained by one input.
FIG. 5 is a diagram illustrating the size of a fingerprint input unit according to the second embodiment.
FIG. 6A is an explanatory diagram of a feature point, and FIG. 6B is an explanatory diagram of a feature point and a relation.
FIG. 7 is a flowchart illustrating the operation of the conventional technique.
[Explanation of symbols]
REFERENCE SIGNS LIST 1 input unit 2 storage device 21 input image storage unit 22 reconstructed image storage unit 23 feature data storage unit 3 data processing unit 31 coordinate calculation unit 32 reconfiguration unit 33 feature data extraction unit 4 collation / authentication device

Claims (6)

An input section having an area smaller than the input object;
An input image storage unit for storing image information obtained from the input unit,
A reconstructed image storage unit for storing image information reconstructed from a plurality of input images,
A feature data storage unit that stores feature data obtained by analyzing the reconstructed image,
When a biometric pattern is input to the input unit, a biometric pattern image for one input is stored in the input image storage unit, and a biometric pattern image stored in the input image storage unit is stored in the reconstructed image storage unit. Compared with the reconstructed image information is calculated position coordinates on the reconstructed image storage unit to be taken by the biological pattern image of the input image storage unit, based on the calculated coordinates, for one input The operation until the biological pattern image is stored in the reconstructed image storage unit is repeated a plurality of times while shifting the input object, and when the reconstructed image reaches a predetermined size, the reconstructed image is Data processing means for analyzing and extracting feature data and storing the feature data in the feature data storage unit;
A biometric pattern information input / verification device, comprising: a verification / authentication unit that determines whether or not the biometric pattern information is registered using the stored feature data. 2. The biometric pattern information input / collation apparatus according to claim 1, wherein the data processing unit discards the reconstructed image for which the characteristic data extraction processing has been completed. The data processing means may determine whether or not a biological pattern has been input to the input unit after completion of the feature data extraction process, and if so, repeat image acquisition and feature data extraction. The biometric pattern information input / collation device according to claim 1 or 2. An input unit having an area smaller than the input object, an input image storage unit that stores image information obtained from the input unit, and a reconstructed image storage unit that stores image information reconstructed from a plurality of input images. A feature data storage unit that stores feature data obtained by analyzing a reconstructed image, and a method of inputting and matching biometric pattern information in an apparatus including:
When a biological pattern is input to the input unit, a biological pattern image for one input is stored in the input image storage unit,
The biometric pattern image stored in the input image storage unit is compared with reconstructed image information stored in the reconstructed image storage unit, and the reconstructed image storage unit to be taken by the biometric pattern image in the input image storage unit Calculate the upper position coordinates,
Based on the calculated coordinates, the operation until storing the biometric pattern image for one input in the reconstructed image storage unit is repeated a plurality of times while shifting the input object,
If the reconstructed image has reached a predetermined size, analyze the reconstructed image to extract feature data, store it in the feature data storage unit,
A method for inputting and matching biometric pattern information, comprising determining whether or not the stored feature data matches registered biometric pattern information. 5. The method for inputting and matching biometric pattern information according to claim 4, wherein the reconstructed image for which the characteristic data extraction processing has been completed is discarded. 6. After completion of the feature data extraction process, it is determined whether or not a biological pattern has been input to the input unit, and if so, image acquisition and feature data extraction are repeated. Input / collation method of biological pattern information.

Images