JPH07271977A - Fingerprint collation processor - Google Patents

Abstract

PURPOSE:To comprehensively shorten processing time by collating plural of collation side fingerprint data with the same file side fingerprint data in parallel to reduce the number of times reading the file side fingerprint data from a data base as much as possible. CONSTITUTION:In order to reduce the number of times reading file side fingerprint data from a file side fingerprint data base 112, each fingerprint data of the same range information group is inputted as the collation fingerprint side data of each collation mechanism 106 to 109, the same data Fi as the file side fingerprint data of each collation mechanism 106 to 109 is inputted in all the collation mechanisms 106 to 109 in parallel and collations are performed. Thus, processing time can be comprehensively shortened.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fingerprint collation processing device, and more particularly to a fingerprint collation processing device for performing parallel processing using a plurality of processing mechanisms.

[0002]

2. Description of the Related Art FIG. 2 is a block diagram showing a conventional processing apparatus of this type. Information about a plurality of fingerprints to be collated is stored in a collated fingerprint file 103 and a plurality of fingerprints are stored. For each known fingerprint, the information of each fingerprint is stored in the file side fingerprint database 112, and the information of one fingerprint in the collation fingerprint file 103 is compared with the information of one fingerprint in the file side fingerprint database 112. Matching mechanisms 106, 107, ... 1 that determine the degree of similarity
The fingerprint collation device 204 is configured by providing a predetermined number of 08 and 109. Then, the matching result by the matching mechanism is stored in the candidate fingerprint file 113 via the matching result control unit 111.

The range of fingerprint information in the file side database 112 to be compared with each fingerprint information in the collation fingerprint file 103 is the collation range setting means 10.
It is set by 2. The collation side data control unit 205
Data from the collation fingerprint file 103 is sequentially read out according to an instruction from the collation range setting means 102, and is input as collation fingerprint side data of each collation mechanism 106, 107, ... 108, 109. Similarly, the file side data control unit 2
10 sequentially reads the data from the file-side fingerprint database 112 according to an instruction from the collation-side data control unit 205, and each collation mechanism 106, 107, ... 108, 10
Input as file side fingerprint data of No. 9.

The collation is carried out in parallel by a plurality of collating mechanisms, but there are M pieces of fingerprint data to be collated and a collation range (the collation range varies depending on the fingerprint data to be collated, but for simplicity of explanation. If the number of fingerprint data (assuming the same for each target data) is N, then M ×
Matching N times is required. In the conventional apparatus, as shown in FIG. 2, one target fingerprint data S _{i is} collated with a collation range N, and this is repeatedly performed for M target fingerprint data.

[0005]

The conventional fingerprint collation processing apparatus shown in FIG. 2 is an efficient collation apparatus when the total number (M) of fingerprint data to be collated is small, and particularly when M = 1. In the parallel processing in this case, the collation can be efficiently performed, but the collation efficiency becomes worse as M increases. That is, it takes a relatively long time to read out the desired file-side data from the file-side fingerprint database 112 and input this to the desired collation mechanism, but in the device shown in FIG. This is because it is necessary to read all file-side fingerprint data within the collation range for one item.

Further, in the conventional apparatus, the two kinds of operations of reading from the database and collating the data are not executed by the pipeline system, so that there is a problem that the overall processing efficiency is poor.

An object of the present invention is to solve these problems in the conventional one and to provide a collation processing device having a high processing efficiency as a whole.

[0008]

The fingerprint collation processing apparatus according to the present invention uses a plurality of collation side fingerprint data for the same file side fingerprint data in order to reduce the number of times of reading the file side fingerprint data from the database as much as possible. We decided to collate in parallel. In addition, reading from the database and collation are processed in parallel by the pipeline method.

That is, in one embodiment of the present invention, a collation fingerprint file that stores information of each fingerprint for a plurality of fingerprints to be collated, and a file side that stores information of each fingerprint of a plurality of known fingerprints. A fingerprint database, a fingerprint collation device provided with a predetermined number of collation mechanisms for comparing the information of one fingerprint in the collation fingerprint file with the information of one fingerprint in the file side fingerprint database, and determining the degree of similarity;
Collating fingerprint setting means for setting a range of fingerprint information in the file side database (hereinafter referred to as collating range) to be compared with each fingerprint information in the collating fingerprint file, provided in the fingerprint collating device, and collating fingerprints Among the fingerprint information in the file, fingerprint information that belongs to the same collation range (referred to as the same range information group) is collated by a plurality of collation mechanisms Collation that sequentially inputs the same data as fingerprint side data so that the same data is not simultaneously input to a plurality of collation mechanisms The side data control unit is provided in the fingerprint collation device, and the data within the collation range of the fingerprint information in the file side fingerprint database is input to the plural collation mechanisms at the same time as the file side fingerprint data of the plural collation mechanisms. File side data control unit that sequentially inputs while allowing When the collation work with the fingerprint data is completed, the collation side fingerprint data or the file side fingerprint data is updated in parallel with the progress of the collation work in other collation mechanism, and each fingerprint data in the same range information group is updated. Is provided with pipeline type control means for controlling the collation side data control section and the file side data control section so that all the fingerprint data within the collation range are collated.

In the pipeline type control means, in the initial state, each fingerprint data of the same range information group is input as the collation fingerprint side data of each collation mechanism,
The same data (provisionally F _i ) is input in parallel to all the collation mechanisms as the file side fingerprint data of each collation mechanism,
Until all the information in the same range information group is input to the matching mechanism, only the matching fingerprint side data of the information mechanism for which the matching work has been completed is updated, and all the information in the same range information group is input to the matching mechanism. the termination information mechanism collation work after the end is left as it collation fingerprint side data, and updates the file side fingerprint data from the F _i to F _{i + 1.}

Further, the file-side data control section includes a buffer memory for storing at least one fingerprint of the file-side fingerprint data, and F _{i is} used as the file-side fingerprint data.
After inputting to all the collation mechanisms, the file side fingerprint data F _{i + 1} to be input next is read from the file side database and held.

Further, in the pipeline system control means, each fingerprint data of the same range information group is divided into each division of the total number of plural collation mechanisms and their fractions, and the pipeline system control means is provided for each divided division. It is characterized by executing.

Further, the file side data control unit is
A buffer memory for storing at least one fingerprint of the file side fingerprint data is provided, and the F side is used as the file side fingerprint data.
After inputting _i to all the collation mechanisms, the file side fingerprint data F _{i + 1} to be input next is read out from the file side database and held.

[0014]

Embodiments of the present invention will now be described with reference to the drawings. 1 is a block diagram showing an embodiment of the present invention, in which the same reference numerals as those in FIG. 2 indicate the same or corresponding parts, and the fingerprint collation device 104 of the present invention is different from the conventional device shown in FIG. The collation data control unit 105 and the file side data control unit 110 of the present invention are the same as the conventional collation side data control unit 2
05, unlike the conventional file side data control unit 210,
The data read control to the collation mechanisms 106, 107, ..., 108, 109 is different, and the file side data control unit 110 is provided with a buffer memory having a capacity of at least one data, and the control by the pipeline method is performed. The point is that it improves efficiency.

FIG. 3 is a flow chart showing the process of matching in the present invention. The same reference numerals as those in FIG. 1 indicate the same parts, and the alphabet S is the data in the same range information group in the matching fingerprint file 103, the alphabet. F is data within the collation range in the file side fingerprint database 112, and the data type is distinguished by the suffix attached to the data. In the example shown in FIG. 3, the case where the data S is 18 pieces of S _{1 to} S ₁₈ and the data F is k pieces of F _{1 to} F _k will be described.

In this case, the collation is started from an initial state indicated by 301 in FIG. That is, different data S _{1 to} S ₁₆ are input as the collation fingerprint side data of a plurality of collation mechanisms, and the same data F ₁ are all input as the file side fingerprint data. When the data F ₁ is input to all the collating mechanisms, the file side data control unit 1
10 reads F ₂ from the file side fingerprint database 112 and inputs it to the buffer memory. Although F ₁ was held in this buffer memory, after F ₁ was input to all the collation mechanisms, it became unnecessary to use F ₁ ,
Replace with F ₂ . It is assumed that collation is performed in parallel with such reading and collation of S ₃ and F ₁ is completed. At this point, the input of the data S to all the collating mechanisms has not been completed yet, and S ₃ is changed to S ₁₇ without changing F _{1 and} the state of 302 in FIG. 3 is obtained.

Next, when the comparison between S ₅ and F ₁ is completed,
S ₅ is changed to S ₁₈ , and the state of 303 in FIG. 3 is obtained. Three
If the collation of S ₂ and F ₁ is completed in the state of 03, all the S data corresponding to F ₁ has already been input to the collating mechanism, so S ₂ is left as it is and the F of the other party is left. Change the data to F ₂ . At this time, since F ₂ has already been stored in the buffer memory, this change is made in a short time and becomes state 304. In this way S ₁ ~
When the matching with F ₁ for all of S ₁₈ is completed, the state 3
It looks like 05. At this point, F ₃ is read from the database 112 and is input to the buffer memory of the file side data control unit 110.

After this, the states 306, 307, 308,
The control is advanced as in ..., and all the S data and the F data are collated. As can be understood from the above description, the collation and the reading from the database 112 are performed by the pipeline control, so that the reading time from the database 112 hardly affects the entire processing time.

The embodiment shown in FIG. 3 shows the case where the pipeline method is used as effectively as possible, but the total processing time can be greatly shortened as compared with the conventional apparatus by a slightly simpler control. For example, the total number of S data is 2
In case of 8, this is 1 division of 16 data and 1 of 12 data
It may be divided into sections and the pipeline system control may be performed for each of the divided sections. FIG. 4 shows a state progress in the case of collating the first division of 16 data.

The state 401 is apparently the same as the state 301 of FIG. 3, but is different in that the collation fingerprints S ₁₇ , S ₁₈ waiting for collation do not exist. Therefore, when the input of the state 401 is completed, the data F ₂ is transferred from the file side fingerprint database 112 to the file side data control unit 1.
It can be read out to 10 buffer memories.
Then, for example, when the comparison between S ₃ and F ₁ is completed, the state 402 is set, and the comparison between S ₃ and F ₂ can be started immediately. In this way, the collated fingerprint data for which collation with F ₁ is completed is gradually increased, and the state 40 and the state 40 are passed.
4, the data F ₃ may be read from the file side fingerprint database 112 into the buffer memory of the file side data control unit 110 when the data F ₂ has been input to all the collation mechanisms.

In this way, when the collation for one section of 16 data is completed, the collation for one section of 12 data is executed next. In this case, only 12 of the 16 matching mechanisms would be used. 3 and 4
As is clear from a comparison of the above, in the embodiment shown in FIG. 4, the effect of the pipeline type control is lower than in the embodiment shown in FIG. 3, but the control becomes simpler. Further, as is clear from a comparison between the conventional apparatus shown in FIG. 2 and the apparatus of the present invention shown in FIG. The device of the invention can reduce the processing time.

[0022]

As described above, the fingerprint collation processing device of the present invention can shorten the total processing time.
In particular, when there are many file-side fingerprint data to be collated, and when there are relatively many data having the same collation range in the collated fingerprint data, the overall processing time can be shortened. In such a case, in the conventional apparatus, it was necessary to read the database once for each piece of fingerprint data to be collated, but in the apparatus of the present invention, it is sufficient to read the database once for every 16 pieces of fingerprint data to be collated. Collation time is shorter than database read time,
Moreover, this reading can be executed in a pipeline manner, and the overall processing time can be shortened.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a block diagram showing a conventional device.

FIG. 3 is a flow chart showing an embodiment of the operation of the apparatus of FIG.

4 is a flow chart showing another embodiment of the operation of the apparatus of FIG.

[Explanation of symbols]

 102 collation range setting means 103 collation fingerprint file 104 fingerprint collation device 105 collation side data control unit 106 collation mechanism 107 collation mechanism 108 collation mechanism 109 collation mechanism 110 file side data control unit 111 collation result control unit 112 file side fingerprint database 113 candidate fingerprint File

Claims (5)

[Claims] 1. A collation fingerprint file that stores information of each fingerprint for a plurality of fingerprints to be collated, a file-side fingerprint database that stores information of each fingerprint of a plurality of known fingerprints, among the collation fingerprint files. A fingerprint collation device provided with a predetermined number of collation mechanisms for comparing the information of one fingerprint of the above and the information of one fingerprint in the file side fingerprint database to determine the degree of similarity, Collation range setting means for setting a range of fingerprint information in the file-side database (hereinafter referred to as collation range) to be compared with respect to fingerprint information, provided in the fingerprint collation device, and included in the collated fingerprint file. Of the fingerprint information, fingerprint information belonging to the same matching range (hereinafter referred to as the same range information group) is used as the same fingerprint side data of the plurality of matching mechanisms as the same fingerprint side data. Data control unit for sequentially inputting data so as not to be input to a plurality of matching mechanisms at the same time, the fingerprint matching device being provided in the fingerprint matching device, and the plurality of data within the matching range among the fingerprint information in the file fingerprint database. A file-side data control unit that sequentially inputs the same data as file-side fingerprint data of the matching mechanism while allowing the same data to be simultaneously input to a plurality of matching mechanisms; and the matching-side fingerprint data and the file among the plurality of matching mechanisms. When the collation work with the side fingerprint data is completed, the collation side fingerprint data or the file side fingerprint data is updated in parallel with the progress of the collation work in the other collation mechanism, and in the same range information group. Of each fingerprint data in the collation range is collated with all the fingerprint data in the collation range. Fingerprint verification processing apparatus characterized by comprising pipeline system control means for controlling the motor control unit. 2. The pipeline type control means inputs each fingerprint data of the same range information group as collation fingerprint side data of each collation mechanism in an initial state,
The same data (provisionally F _i ) is input in parallel to all the collation mechanisms as the file side fingerprint data of each of the collation mechanisms until all the information in the same range information group is input to the collation mechanisms. Update only the collation fingerprint side data of the information mechanism for which the collation work has been completed, and the collation fingerprint side for the information mechanism for which the collation work has been completed after all the information within the same range information group has been input to the collation mechanism. the data kept unchanged, F _{i + 1} file side fingerprint data from the F _i
The fingerprint collation processing device according to claim 1, wherein the fingerprint collation processing device is updated. 3. The file-side data control section comprises a buffer memory for storing at least one fingerprint of the file-side fingerprint data, and after inputting F _i as the file-side fingerprint data to all the collation mechanisms, 3. The fingerprint collation processing apparatus according to claim 2, wherein the file side fingerprint data F _{i + 1} to be input to is read out from the file side database and held. 4. The pipeline type control means divides each fingerprint data of the same range information group into respective divisions of the total number of the plurality of collation mechanisms and their fractions, and the divided 1
The fingerprint collation processing apparatus according to claim 1, wherein the pipeline system control means is executed for each section. 5. The file-side data control unit includes a buffer memory for storing at least one fingerprint of the file-side fingerprint data, and after inputting F _i as the file-side fingerprint data to all the matching mechanisms, 5. The fingerprint collation processing apparatus according to claim 4, wherein the file side fingerprint data F _{i + 1} to be input to is read out from the file side database and held.