JPH11265451A - Image information inquiry system - Google Patents

Abstract

(57) [Summary] [Problem] In an image information inquiry system, processing concentration,
Responsiveness and reliability are problematic due to data concentration. One or a plurality of image input devices for electronically inputting image information and additional information describing the image information, and one for storing and managing the image information and the additional information generated by the image input device. One or more image management devices, and one or more image information and additional information generated by the image input device, registered in the image management device, and compared with the image information and the additional information stored in the image management device. An image collating device and one or more information inquiry devices for inquiring image information and additional information stored in the image management device using the image management device and the image collating device are connected via a network.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image information inquiry system using a computer, and more particularly to an image information inquiry system in a distributed network environment aiming at improving retrieval efficiency by load distribution and improving reliability by risk distribution. .

[0002]

2. Description of the Related Art As a conventional image information inquiry system,
"Police Information Management System" (Information Processing Society of Japan, Information Processing Society of Japan 57-2, January 16, 1996) and "Personal Identification Technology Using Fingerprints and Palm Prints" (Information Processing Society of Japan, Vol.25, No.6) , June 1984). According to this document, the fingerprint automatic identification system is a system that answers a personal identification result to an inquiry by a police fingerprint image.

[0003] The Japanese police organization forms a pyramid-type organization with the National Police Agency at the top, a regional police bureau, a prefectural police headquarters, and a police station underneath. Since then
The prefectural police headquarters is called the prefectural police headquarters, and the police station is called the station.

An automatic fingerprint identification system is installed at the National Police Agency, and is used to electronically print fingerprint base paper sent from the whole country by mail or the like, which is stamped with ink at a station or the prefectural police headquarters, or a latent fingerprint collected at the site. Processing such as reading, image correction, feature amount extraction, registration, or matching of fingerprint patterns having similar feature amounts is concentrated. Data is also centrally managed by the National Police Agency.

[0005]

In the above fingerprint automatic identification system, electronic reading, image correction, feature extraction,
Processing such as registration and verification is concentrated. Since the processing is concentrated in one place, it takes time to obtain the query result. In addition, since data is concentrated in one place, the risk in the event of a disaster is large, and the responsiveness when access is concentrated is poor. Therefore, an object of the present invention is to realize a reduction in response time and an improvement in reliability by processing distribution and data distribution in an image information inquiry system. Other than the police, there is a similar problem in the certificate authority system in electronic commerce.

[0006]

Means for Solving the Problems To solve the above problems, the following means are provided. One or more image input devices for electronically inputting image information and additional information describing the image information, and one or more image storage devices for storing and managing the image information and additional information generated by the image input device. An image management device, and one or more image matching devices for registering the image information and the additional information generated by the image input device with the image management device and checking the image information and the additional information stored in the image management device. Then, one or a plurality of information inquiry devices for inquiring image information and additional information stored in the image management device using the image management device and the image collating device are connected via a network.

The image input device may have a feature extraction support function for extracting a feature that characterizes the image information. Further, the image management device has a hierarchical archive according to the generation time of the image information that is stored and managed, and stores the image information of a relatively old time in a large-capacity storage device and the image information of a relatively new time in a high-speed storage device. , Manage hierarchically. Alternatively, the image management device has a hierarchical archive according to the amount of image information or additional information that is stored and managed,
The image information may be hierarchically managed in the mass storage device, and the additional information may be managed in the high-speed storage device.

Further, when registering or inquiring image information or additional information using an image input device or an information inquiry device, the same or different image collating devices can be registered in different image management devices until they can be registered and collated in an appropriate image management device. The device may be accessed. Further, the image information is compressed, and the additional information is processed without compression. As described above, data management, registration, and inquiry can be made more efficient, response time can be shortened, and risks can be avoided.

[0009]

FIG. 1 is a block diagram showing an embodiment of the present invention.
Will be described in detail with reference to FIG. The present embodiment is a fingerprint inquiry registration system that supports management, collation, and registration of fingerprints in a police criminal investigation.

FIG. 1 shows an overall system configuration for realizing this embodiment. The National Police Agency system 101 has DAFIS
(Distributed Automated Fi
ngerprint Identification
System) 102, FIMS (Fingerpri
nt Information Management
A system 103 and an information inquiry system 104 are connected by a LAN (Local Area Network). Prefectural Police Headquarters System 105 has ILS
(Intelligent Live Scanne
r) 106, DAFIS 102, FIMS 103 and information inquiry system 104 are connected by LAN. The ILS 106 and the information inquiry system 104 are connected to the station system 107 via a LAN. In addition, the National Police Agency system 101 and the prefectural police headquarters system 105, and the prefectural police headquarters system 105 and the station system 107 are WAN
(Wide Area Network). The DAFIS 102 is a device in which the collation registration device is made independent and distributed to regions (for example, prefectural police headquarters). As a result, a bottleneck due to the concentration of the processing load can be avoided.

The FIMS 103 has a server configuration in consideration of the Internet / intranet correspondence. In addition, by distributing to regions (for example, prefectural police headquarters), DA
The processing load is dispersed by the FIS 102 and a bottleneck is avoided. The FIMS 103 has a hierarchical archive 108
With a relatively old fingerprint image in MODL (Mag
neto Optical Disk Liblar
The fingerprint image of a relatively new age is hierarchically managed in a high-speed storage device such as a disk array in a mass storage device such as y). Alternatively, a fingerprint image is hierarchically managed on a high-capacity storage device such as a disk array, and additional information such as a reference number, a finger type, and a pattern, including a feature amount for specifying a fingerprint, is stored in a high-capacity storage device such as a MODL. . In addition, WWW
(World Wide Web) server 109 is provided,
The connection with the Web may be achieved.

The information inquiry system 104 is a system for efficiently searching for fingerprint images and additional information that are distributed and managed on a network. Using the WWW browser, WW
The connection with the W server 109 may be achieved.

The ILS 106 is a live scanner 110
In addition, a fingerprint image pre-processing 111 and a feature amount extraction support 112 function are provided and distributed to regions (for example, prefectural police headquarters and departments), thereby dispersing and reducing the processing load. The live scanner 110 is a device that optically reads an imprint fingerprint and electronically records it. The preprocessing 111 is a function of converting a fingerprint image obtained by the live scanner 110 into an image that can be used for the feature amount extraction support 112. Feature extraction support 1
Reference numeral 12 denotes a function for extracting a feature amount characterizing the fingerprint image from the image converted in the preprocessing 111.

In the fingerprint inquiry registration system, the ILS 106
Alternatively, when registering or inquiring a fingerprint image or additional information using the information inquiry system 104, an appropriate F
The same or different D until it can be registered and verified in IMS
AFIS accesses different FIMS. Thus, registration or inquiry can be made without being aware of the location of the appropriate FIMS. In addition, a fingerprint image having a relatively large amount of information is compressed, and additional information having a relatively small amount of information is not compressed. As a result, the network load can be reduced and the storage device can be saved for the fingerprint image, and the decompression time is not required for the additional information, so that high-speed collation can be performed.

Although FIG. 1 does not show the district police station, it can be considered in the same way as the others. In addition, the prefectural police headquarters system 105 and the station system 107 can be expanded to a plurality. Furthermore, DAFIS102, FIM
S103, the information inquiry system 104, and the ILS 106 can each be extended to a plurality. The FIMS 103 may store data of the same content in a plurality of different FIMSs. Thereby, risk distribution and access distribution can be achieved.

FIG. 2 is a system configuration diagram focusing on the device configuration of the present embodiment. Office 201 and Prefectural Police Headquarters 20
2 includes a live scanner 110 and a feature amount extraction device 203.
And an information search device 204. Department 201
The prefectural police headquarters 202 exchanges data via the communication line 205. For connection to the communication line 205, TA (Te
(Rminal Adapter) 206 is used. Further, the encryption device 207 may be installed as needed.
The feature amount extraction device 203 is a computer that performs pre-processing 111 and feature amount extraction support 112, and includes a live scanner 110.
Are connected to the TA 206 or the encryption device 207. The information search device 204 is a computer that implements the information inquiry system 104, and includes the TA 206 or the encryption device 2
07. The feature amount extraction device 203 and the information retrieval device 204 may be realized by the same computer.

The district police station 208 is provided with a fingerprint collation processing unit 209, a data management unit 210, a feature amount database 211 of the own division, an image database 212 of the own division, and a personal database 213 of the own division. . Prefectural Police Headquarters 202 and District Police Bureau 208
4 to transfer data. The fingerprint matching processing device 209 is a computer that realizes the DAFIS 102,
It is connected to the data management device 210. The data management device 210 is a computer that implements the FIMS 103,
A206, fingerprint collation processing device 209, feature amount database 211, image database 212, and personal database 213 are connected. Feature database 211
Is a storage device for storing feature amount data of the own management division, and is connected to the data management device 210. The feature amount data is a feature amount extracted by the feature amount extraction device 203, and is stored in, for example, a disk array because the information amount is relatively small. The image database 212 is a storage device for storing image data of the own section, and is connected to the data management device 210. Since the image data is a fingerprint image and has a relatively large information amount, it is stored in, for example, a MODL. The personal database 213 is a storage device for storing personal data of the own management division, and the data management device 21
0 is connected. The personal data is an identity, a criminal record, and the like, and since the amount of information is relatively small, it is stored in, for example, a disk array.

The National Police Agency 215 has a fingerprint collation processor 20
9, a data management device 210, a feature database 216 for the whole country, an image database 217 for the whole country, and a personal database 218 for the whole country. The district police station 208 and the National Police Agency 215 exchange data via the communication line 219. The fingerprint collation processing device 209 uses D
This is a computer that implements the AFIS 102 and is connected to the data management device 210. The data management device 210
A computer that implements the IMS 103, and is connected to the TA 206, the fingerprint matching processing device 209, the feature database 216, the image database 217, and the personal database 218. The feature database 216, the image database 217, and the personal database 218 are storage devices similar to the feature database 211, the image database 212, and the personal database 213, respectively, and store data for the whole country.

Station 201, Prefectural Police Headquarters 202, District Police Bureau 2
08 can be extended to a plurality. In addition, live scanner 1
10 or the feature extraction device 203 or the information retrieval device 2
04 may be installed in the district police station 208 or the National Police Agency 215, or the fingerprint matching processing device 209, the data management device 210, or the local database may be
Or the prefectural police headquarters 202. Furthermore, each device can be locally expanded to a plurality.

In the fingerprint inquiry registration system, consistency of data stored in a plurality of layers is maintained. For example, in FIG. 2, the feature database 216 for the whole country, the image database 217 for the whole country, and the personal database 218 for the whole country are respectively the feature database 211 of the own section, the image database 212 of the own section, and the Personal database 213
And that the same data is stored for the own division. FIG. 6 is a flowchart of a data matching process for maintaining data consistency. First Regional Police Department 208
The new or updated data transmitted to the server is stored in the database of the own section (601). One ID is assigned to one person in the database of the own division. That is, the same ID is set for the feature amount data, image data, and personal data of the same person. If the transmitted data is new data (602), an ID that is not used in the database of the own division is set (603), and the ID and the new data are sent to the TA of the regional police station 208 (604). . If it is update data (602), the current ID and the update data are sent to the TA of the regional police station 208 (604). The TA of the district police station 208 adds information for specifying the source, for example, an IP address to the ID (605), and transmits it to the National Police Agency 215 in the data structure shown in FIG. 7 (606). ID701 has information added to identify the sender
ID. The new feature data 702, new image data 703, and new personal data 704 are new or updated feature data, image data, and personal data, respectively. If there is no corresponding data, it is set to null (information indicating that no data exists). These data are stored in a database for the whole country using ID701 (607). As described above, it is possible to perform decentralization while maintaining data consistency.

In the fingerprint inquiry registration system, consistency of data stored in a plurality of layers is maintained. For example, in FIG. 2, the feature database 216 for the whole country, the image database 217 for the whole country, and the personal database 218 for the whole country are respectively the feature database 211 of the own section, the image database 212 of the own section, and the Personal database 213
And that the same data is stored for the own division. FIG. 6 is a flowchart of a data matching process for maintaining data consistency. First Regional Police Department 208
The new or updated data transmitted to the server is stored in the database of the own section (601). One ID is assigned to one person in the database of the own division. That is, the same ID is set for the feature amount data, image data, and personal data of the same person. If the transmitted data is new data (602), an ID that is not used in the database of the own division is set (603), and the ID and the new data are sent to the TA of the regional police station 208 (604). . If it is update data (602), the current ID and the update data are sent to the TA of the regional police station 208 (604). The TA of the district police station 208 adds information for specifying the source, for example, an IP address to the ID (605), and transmits it to the National Police Agency 215 in the data structure shown in FIG. 7 (606). ID701 has information added to identify the sender
ID. The new feature data 702, new image data 703, and new personal data 704 are new or updated feature data, image data, and personal data, respectively. If there is no corresponding data, it is set to null (information indicating that no data exists). These data are stored in a database for the whole country using ID701 (607). As described above, it is possible to perform decentralization while maintaining data consistency.

In the fingerprint inquiry registration system, the ILS 106
When matching feature values using, an appropriate FIMS is accessed. FIG. 3 is a flowchart of the distributed matching process. First, feature amount data is extracted with feature amount extraction support (112). The feature data is sent to the DAFIS of the local system via the local network.
The DAFIS accesses the FIMS in the same system and performs collation in the local system (30).
1). Here, the term “local” means being physically close to each other, and corresponds to the own pipe division in FIG. It is highly possible that the fingerprint of the person to be collated is registered in the local system FIMS. Therefore, processing load and network load are not applied.

As a result of collation in the local system, if there is matching data, the collation result is returned to the ILS 106, and if there is no matching data, collation is performed in another system (302). When all the systems have been searched, the matching result is returned to the ILS 106 (303). If the entire system has not been searched yet, the feature data is sent to DAFIS of another system via the WAN. The DAFIS accesses the FIMS in the same system and performs collation in another system (304). The other here is the
This is equivalent to the whole country. In general, DAFIS in the same system as the FIMS to be accessed is used. However, when DAFIS in the same system cannot be used, DAFIS of a different system may be used.

In the distributed matching process of FIG. 3, the function of the feature amount extraction support 112 is dispersed, so that only feature amount data having a relatively small information amount needs to flow through the network, so that the network load is small.

FIG. 4 is a block diagram of an inked fingerprint inquiry in the fingerprint inquiry registration system. The purpose of the imprint fingerprint inquiry is to confirm the identity and criminal history of the arrested suspect using the imprint fingerprint. Even if the suspect misrepresents his name or remains silent, the fact can be promptly revealed by the imprint fingerprint inquiry.

First, at the office or the prefectural police headquarters, the suspect is fingerprint-printed 401 using a live scanner.
The live scanner generates imprint fingerprint image data 402. Pre-processing 111 such as thinning is performed on the imprinted fingerprint image data 402 to convert it into feature amount extraction data 403. From the feature amount extraction data 403, impression fingerprint feature amount data 405 characterizing the impression fingerprint image data 402 is extracted by feature amount extraction 404. The above is IL
S (Intelligent Live Scanne)
r) The function of 106. The output of ILS 106 is
Imprint fingerprint image data 4 generated by the fingerprint imprint 401
02 and the impression fingerprint characteristic amount data 405 extracted by the characteristic amount extraction 404, which are compressed or encrypted and transmitted to the district police station or the National Police Agency.

At the regional police station or the National Police Agency, the imprint fingerprint feature data 405 transmitted from the station or the prefectural police headquarters is collated 408 with all the feature data 407 stored in the feature database 406. Collation 40
The imprinted fingerprint feature amount data 405 used for No. 8 is stored in the feature amount database 406, and the imprinted fingerprint image data 402 transmitted from the signature or the prefectural police headquarters is stored in the image database 409.
Is registered 410.

As a result of the comparison 408, a candidate fingerprint similar to the reference fingerprint is output from the image database 409 as candidate fingerprint image data 411. The final confirmation 412 of whether or not the candidate fingerprint matches the reference fingerprint is performed by the naked eye using the candidate fingerprint image data 411 and the impression fingerprint image data 402. Visual confirmation 412
As a result, personal data 413 such as identity and criminal history, which is additional information of the matching candidate fingerprint, is stored in a personal database 414.
And send it to the station or prefectural police headquarters. The station or the prefectural police headquarters receives the result 415 of the transmitted personal data 413. All of the above may be processed online.

FIG. 5 is a block diagram of a latent fingerprint inquiry in the fingerprint inquiry registration system. The purpose of the latent fingerprint inquiry is to use fingerprints (latent fingerprints) collected from crime scenes,
It is to find the culprit in the case.

First, at a department or the prefectural police headquarters, the latent fingerprint obtained by the on-site fingerprint collection 501 is digitized with high accuracy using a camera or a scanner. As a result, latent fingerprint image data 502 is obtained. Since the collected latent fingerprints include those of a person other than the culprit (herein referred to as a related person), it is necessary to perform 503 (screening) of the related fingerprint of the latent fingerprint. Unnecessary latent fingerprints (not those of the criminal, that is, those of the parties involved) can be shaken off from the plurality of latent fingerprints by collating the imprint fingerprints of the parties with the latent fingerprints collected at the site. Therefore, the related person fingerprint imprint 504 is performed using a live scanner. The live scanner generates imprint fingerprint image data 402. The imprint fingerprint image data 402 and the latent fingerprint image data 502
503 (screening) is performed by collation and confirmation using ILS in FIG.
(Intelligent Live Scanne
r) 106 is a live scanner that has a high added value by adding a screening function of related person fingerprints, such as preprocessing, trace support, feature amount extraction, collation, and confirmation support. The output of the ILS 106 is two types of data about the unknown latent fingerprint that is not the latent fingerprint of the person concerned. One is the unknown latent fingerprint feature data 505 extracted for collation in the process of the related party drop-off 503, and the other is the unknown latent fingerprint image data 506 remaining as a result of the related party drop-off 503. To the district police station or the National Police Agency after compression or encryption.

In the regional police station or the National Police Agency, the unknown latent fingerprint feature data 505 transmitted from the station or the prefectural police headquarters is collated 408 with all the feature data 407 stored in the feature database 406. The unknown latent fingerprint feature amount data 505 used for the matching 408 is registered 410 in the feature amount database 406, and the unknown latent fingerprint image data 506 transmitted from the signature or the prefectural police headquarters is registered 410 in the image database 409. As a result of the comparison 408, a candidate fingerprint that is similar to the reference fingerprint is output from the image database 409 as candidate fingerprint image data 411. The final confirmation 412 as to whether or not the candidate fingerprint matches the reference fingerprint is performed by the naked eye using the candidate fingerprint image data 411 and the unknown latent fingerprint image data 506. As a result of the confirmation 412 by the naked eye, personal data 413 such as identity and criminal history, which is additional information of the matching candidate fingerprint
Is retrieved from the personal database 414 and transmitted to the department or the prefectural police headquarters. The station or the prefectural police headquarters receives the result 415 of the transmitted personal data 413. All of the above may be processed online.

The divisions of the station, the prefectural police headquarters, the regional police bureau, and the National Police Agency in the block diagrams shown in FIGS. 4 and 5 are not limited to those described above. This embodiment is a fingerprint inquiry registration system characterized by load distribution, data distribution, and added value of live scanners. As a result, it is possible to shorten the response time, improve the reliability, and improve the work efficiency.

The above is the detailed embodiment of the present invention. The present invention may be used for a system for sharing not only fingerprints but also various crime information, investigation information, and police administration information distributed throughout the country via a network. . It may be used in a system for specifying the vehicle type from the tire trace image in the police.

The present invention can be widely applied not only to a police system but also to an individual identification system such as a certificate authority in electronic commerce, an individual identification system, an information management search system, and the like.

[0035]

According to the present invention, in an image information inquiry system, load distribution, risk distribution and network load can be reduced, response time can be shortened, reliability can be improved, and data management, registration, and inquiry can be made more efficient. Is possible.

[Brief description of the drawings]

FIG. 1 is an overall system configuration diagram of a fingerprint inquiry registration system.

FIG. 2 is an apparatus configuration diagram of a fingerprint inquiry registration system.

FIG. 3 is a flowchart of a distributed matching process.

FIG. 4 is a block diagram of an imprint fingerprint inquiry using the present invention.

FIG. 5 is a block diagram of a latent fingerprint inquiry using the present invention.

FIG. 6 is a flowchart of a data matching process.

FIG. 7 is a diagram showing a data structure in a data matching process.

[Explanation of symbols]

101: National Police Agency system, 102: DAFIS (Dis
tributed Automated Finger
print Identification System
em), 103 ... FIMS (Fingerprint)
Information Management Sy
stem), 104: Information inquiry system, 105: Prefectural police headquarters system, 106: ILS (Intelligen)
t Live Scanner), 107: signature system, 108: hierarchical archive, 109: WWW (Wor)
ld Wide Web) server, 110: live scanner, 111: preprocessing, 112: feature extraction support

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Seiwa Moto, 890 Kashimada, Saiwai-ku, Kawasaki-shi, Kanagawa 12 Information Systems Division, Hitachi, Ltd. No. 2 Hitachi Engineering Co., Ltd.

Claims (9)

[Claims] 1. One or more image input devices for electronically inputting image information and additional information describing the image information, and storing and managing the image information and the additional information generated by the image input device. One or more image management devices, and the image information and the additional information generated by the image input device are registered in the image management device, and are compared with the image information and the additional information stored in the image management device. One or more image collating devices, and one or more information inquiry devices for inquiring image information and additional information stored in the image management device using the image management device and the image collation device, An image information inquiry system, which is connected via a network. 2. The image information inquiry system according to claim 1, wherein the image input device has a feature amount extraction support function for extracting a feature amount characterizing the image information from the image information. Image information inquiry system. 3. The image information inquiry system according to claim 1, wherein said image input device has a screening function of selecting necessary image information from a plurality of image information. 4. The image information inquiry system according to claim 1, wherein the image management device manages the image information hierarchically in accordance with the time when the image information being stored and managed is generated. An image information inquiry system having an archive. 5. The image information inquiry system according to claim 4, wherein said hierarchical archive stores image information of a relatively old age in a mass storage device and image information of a relatively new age in a high-speed storage device. An image information inquiry system, characterized in that: 6. The image information inquiry system according to claim 1, wherein said image management device is adapted to store said image information or said additional information in a hierarchical manner according to the amount of image information or additional information stored and managed. Image information query system, characterized in that it has a hierarchical archive managed by the system. 7. The image information inquiry system according to claim 6, wherein the hierarchical archive stores the image information in a mass storage device,
An image information inquiry system characterized by storing additional information in a high-speed storage device. 8. The image information inquiry system according to claim 1, wherein when registering or inquiring image information or additional information using an image input device or an information inquiry device, registration and collation can be performed with an appropriate image management device. Up to this point, the same or different image collating apparatus has means for accessing different image management apparatuses. 9. The image information inquiry system according to claim 1, wherein said image information is compressed, and said additional information is uncompressed.