Image analysis – Applications – Personnel identification
Reexamination Certificate
2000-05-11
2001-08-14
Mehta, Bhavesh (Department: 2721)
Image analysis
Applications
Personnel identification
Reexamination Certificate
active
06275601
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method for renewing dictionary image data in a fingerprint identification device, and particularly to a method for renewing dictionary images in a fingerprint identification device, in which dictionary identification time can be shortened.
The present invention relates to a method for renewing dictionary images in a fingerprint identification device, and particularly to a method for renewing dictionary images in a fingerprint identification device, in which the dictionary images can be kept in the best condition.
The present invention relates to a dictionary registration and identification method for a fingerprint identification device, and particularly to a multiple-dictionary registration and identification method for a fingerprint identification device, in which an identification rate is improved by registering and identifying a plurality of dictionaries.
The present invention relates to a method for judging fingerprinting in a fingerprint identification device, and particularly to a fingerprinting judging method in a fingerprint identification device, in which continuous identification rejections can be prevented.
The present invention relates to a fingerprint identification device which can prevent misjudgment in identifying a fingerprint and to a fingerprinting judging method.
2. Description of the Related Art
In a fingerprint identification device, by identifying an entered fingerprint image with fingerprint dictionary data previously registered in a dictionary memory to judge whether or not the entered fingerprint image corresponds to a finger of a particular person whose dictionary data is previously registered in the fingerprint identification device.
In such a case, it is desired that time required for dictionary identification can be shortened by enabling to quickly identify the conformity with the multiple dictionary data to be used for identification.
The fingerprint dictionary data is image data of multiple characteristic parts, such as ending points and bifurcation points, of a fingerprint image extracted by a sensor and registered together with its coordinate data, and when the entered fingerprint image is detected to conform with some of such characteristic data, it is determined to be identification pass.
On the other hand, in fingerprinting by using a sensor, it has been experienced that fingerprints are relatively easily deformed depending on the position of fingers on a fingerprinting surface, the magnitude of pressing a finger against the fingerprinting surface or the like, and the fingerprint image to be taken into the fingerprint identification device is varied.
Therefore, if dictionary data is registered as many as possible on the same person, various cases can be covered even when the fingerprints are deformed as described above. Adopting a method of determining as an identification pass when the entered fingerprint image matches with any one of dictionary images, an identification match rate (acceptance rate) of the fingerprint identification device can be improved.
FIG. 14
shows the configuration of conventional dictionary data, in which (a) indicates all dictionary data, (b) dictionary data on one person within all dictionary data, and (c) one characteristic data in the dictionary data on one person.
All dictionary data comprises dictionary data on M to-be-identified persons (1 to M), for example. Dictionary data on one person comprises N minutiae (minutiae 1 to N), for example. And, one minutia data comprises image data on minutiae and its coordinate data in a prescribed window.
In the fingerprint identification device, by using dictionary data having N minutia data registered therein, pattern matching is made from minutia 1 in order with respect to the entered fingerprint image which is taken in at the fingerprint identification, and if a pattern matches on, for example, L minutia data, it is determined as identification pass.
FIG. 15
shows an example of identifying minutiae in which (a) indicates Example 1, and (b) Example 2. Among N minutiae, those with matched patterns are indicated by ◯, and those without matched patterns by ×.
Example 1 shows that minutia 2 does not match but other minutiae match, and it is determined to be identification pass based on the pattern matching of (L+1) minutiae. And, Example 2 shows that minutiae 2, 3 and 5 do not match but other minutiae match, and it is determined to be identification pass based on the pattern matching of (L+3) minutiae.
Besides, in the fingerprint identification device, to complete the identification with the dictionary data soon, when N minutiae have K minutiae which do not have matched patterns in succession, it is judged that the entered fingerprint image and the dictionary data are not of the same person, it is determined as identification rejection, and identification of the remaining minutiae with the dictionary data is canceled.
FIG. 16
shows the cancellation of the identification process, in which dictionary data comprising N minutiae is identified in succession, K (=3) continuous identification mismatches are detected at minutiae
5
through
7
, and the process is terminated at this point as identification rejection. In this case, even if it is possible that there may be a minutia having a matched pattern among, for example, minutia 8 through minutia N, the remaining process is canceled.
In identifying the minutiae shown in
FIG. 15
, when minutia data that have patterns which often do not match with entered fingerprint images are contained many in minutia data of less than L, time required before patterns are matched with L minutia data becomes longer than when such minutia data are not contained.
And, in the cancellation of the identification process shown in
FIG. 16
, when the minutia data that have patterns which do not match with the entered fingerprint images are contained at least K in succession in the neighborhood of the head of the N minutia data string, an identification match rate (acceptance rate) is lowered.
FIG. 1
shows the configuration of a conventional fingerprint identification device to which the present invention are applied and which comprises a core section
10
and a control
20
. The core section
10
performs registration and identification of fingerprint images, and the control
20
controls the processing in the core section
10
. The control
20
sends a command to the core section
10
and receives from the core section a response to the command.
In the core section
10
, reference numeral
11
denotes a central processing unit (CPU) for controlling the processing of registration and identification of fingerprint images. And, reference numeral
12
denotes a sensor for taking the image of a fingerprint which is to be registered or identified,
13
a frame memory for storing the image taken from the sensor
12
as a multivalued image,
14
an image processing circuit for processing various images, such as binarization of a multivalued image, extraction of minutiae from a binarized image, pattern matching of a binarized image with minutiae and the like,
15
a binary memory for storing a binarized image,
16
a dictionary memory for storing the minutiae extracted from a binarized image as a dictionary image, and
17
an identification memory for keeping the minutiae of fingerprints to be used for identification; and these are connected to a bus
18
of the CPU
11
. And, reference numeral
19
is a control interface for interfacing of a command and a response between the bus
18
and the control
20
.
As a dictionary registration identifying method in such a fingerprint identification device, use of a plurality of dictionary images has been proposed to improve an identification rate.
The dictionary images are minutiae such as ending points and bifurcation points extracted from the image of a fingerprint pressed on the sensor, and when some of such characteristics parts are detected to match with the entered image, it is determined t
Narasaki Koichi
Okazumi Mitsunobu
Soga Takayuki
Souno Hiroyuki
Yamaguchi Masahiko
Blackwell Sanders Peper & Martin
Fujitsu Denso Ltd.
Mehta Bhavesh
LandOfFree
Fingerprinting judging method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Fingerprinting judging method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fingerprinting judging method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2461606