Image analysis – Applications – Personnel identification
Reexamination Certificate
1998-03-23
2001-05-15
Mehta, Bhavesh (Department: 2621)
Image analysis
Applications
Personnel identification
Reexamination Certificate
active
06233348
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a fingerprint registering apparatus, a fingerprint identifying apparatus, and a fingerprint identifying method, in particular, to those suitable for identifying a person.
2. Description of the Related Art
As computers have been widely used in our society, system security has gained public attention. For example, as a means for identifying a person who accesses a computer room or who uses a terminal unit, an ID card or an unique password is used. However, such means have a problem from a viewpoint of security.
Instead of using an ID card or a password, a person identifying technology that uses information intrinsic to a living body is desired. When such information is used for identifying a person, since other people do not have the same information, the information of the living body assures the identity of the person.
As types of information of the living body used for identifying a person, there are fingerprint, voice-print, iris, distribution chart of retinal veins, signature, and so forth. These types of information of a living body are converted into electronic information such as an image by various sensors such as a CCD camera. Thereafter, information of a living body obtained by such a sensor is processed and then information that becomes a key for identifying a person is extracted. Therefore, information of a living body that has been registered is identified with key information of a living body that has been input so as to identify a person.
Next, as an example of information of a living body, a fingerprint will be described.
A fingerprint has two major features that are “no fingerprints are the same” and “no change in a life time”. Thus, it is considered that a fingerprint is the most effective means for identifying a person. Consequently, many simple person-identifying systems using fingerprints have been intensively studied and developed.
A fingertip of a human being has fine concave portions and fine convex portions. A string of a convex portion is referred to as a ridge. Ridges form a pattern intrinsic to a person. When a ridge is traced, a ridge bifurcation and a ridge ending take place. At a ridge bifurcation, the ridge bifurcates out into two ridges. At a ridge ending, the ridge breaks. Since the distribution of ridge bifurcations and ridge endings varies person by person, these points are referred to as feature points. The feature points are used as a prominent means for identifying a person. When fingerprints are identified, it is determined whether the positions, types, and directions of these feature points match.
FIG. 1
is a flowchart showing a conventional fingerprint registering process and a conventional fingerprint identifying process.
In the fingerprint registering process, a fingerprint image to be registered is sampled by a fingerprint sensor (at step S
191
). The sampled fingerprint image is binarized (at step S
192
).
Next, thin lines are generated with the binarized fingerprint image (at step S
193
). Thus, a thin line image of a fingerprint in which the width of each ridge is expressed by one pixel is obtained.
Thereafter, the positions of the feature points of the fingerprint are located and extracted on the thin line image of a fingerprint (at step S
194
). Since the extracted feature points generally contain incorrect feature points, these incorrect feature points are removed (at step S
195
). When these incorrect feature points are removed, if two ridge endings are opposite each other, separated by a short distance, these ridge endings are treated as one ridge that has broken in the image sampling process. In this case, the ridge between the two ridge endings is restored. Thus, the two ridge endings are removed. On the other hand, when two parallel ridges adhere in the middle and thereby a ridge bifurcation takes place, the ridge bifurcations are treated as two parallel ridges that have adhered in the image sampling process. In this case, the ridges that have adhered are separated and thereby the ridge bifurcation is removed.
Next, fingerprint information of each feature point extracted from the fingerprint image is collected (at step S
196
). The fingerprint information is stored as registered fingerprint data in a fingerprint data registering unit
110
.
In the fingerprint identifying process, a fingerprint image to be identified is sampled by a fingerprint sensor (at step S
197
). The sampled fingerprint image is binarized (at step S
198
).
Next, thin lines are generated with the binarized fingerprint image (at step S
199
). Thus, a thin line image of a fingerprint, of which the width of each ridge is expressed by one pixel is obtained.
Next, the positions of features points of the fingerprint are located and extracted on the thin line image of a fingerprint (at step S
200
). Since the extracted feature points generally contain incorrect feature points, the incorrect feature points are removed (at step
5201
).
Next, fingerprint information of each feature point extracted from the fingerprint image is collected (at step S
202
). The registered fingerprint data is read from a fingerprint data registering unit
110
so as to align the positions of the input fingerprint image and the registered fingerprint image (at step S
203
).
Next, the fingerprint information of the input fingerprint image and the fingerprint registered data that has been read from the fingerprint data registering unit
110
are compared so as to identify the input fingerprint image and the registered fingerprint image. In other words, the number of feature points that match in the input fingerprint image and the registered fingerprint image is counted (at step S
204
). When the number of feature points that match in the input fingerprint image and the registered fingerprint image exceeds a predetermined value (at step S
205
), it is determined that the fingerprint of the registered fingerprint image is the same as the fingerprint of the input fingerprint image (at step S
206
). On the other hand, when the number of feature points that match in the input fingerprint image and the registered fingerprint image is smaller than the predetermined value (at step S
205
), it is determined that the fingerprint of the registered fingerprint image is different from the fingerprint of the input fingerprint image (at step S
207
).
As feature point information of feature points, the positions (coordinates), types, and directions of individual feature points are generally used. By determining whether the positions, types, and directions of feature points match in the input fingerprint image and the registered fingerprint image, the fingerprints are identified.
However, when the positions, types, and direction of feature points of the input fingerprint image and the registered fingerprint image are compared, a high identification ratio cannot be obtained. Since the skin of a finger partially expands, shrinks, or rotates, whenever a fingerprint image is sampled, the fingerprint distorts. Thus, the positions and directions of feature points delicately vary. In addition, due to an improper pressure of a finger to the fingerprint sensor, dirty thereof, chaps of a finger, and so forth, whenever a fingerprint image is sampled, a ridge ending and/or a ridge bifurcation is unstably detected. Thus, the identification ratio deteriorates.
SUMMARY OF THE INVENTION
Therefore, an object of the present invention is to provide a fingerprint identifying apparatus and a fingerprint identifying method that allow a fingerprint to be accurately and stably identified.
To solve the above-described problem, according to the present invention, a feature point on a ridge is extracted from fingerprint data. Corresponding to the placement state of the feature point to the ridge, a fingerprint is identified.
Thus, even if a fingerprint distorts due to a partial expansion, shrinkage, rotation, and so forth of the skin of a finger, the placement state of feature point on the ridge does not vary. Thus, a fingerprint can be acc
Fujii Yusaku
Shinzaki Takashi
Yokoyama Ken
Armstrong Westerman Hattori McLeland & Naughton LLP
Fujitsu Limited
Mehta Bhavesh
LandOfFree
Fingerprint registering apparatus, fingerprint identifying... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Fingerprint registering apparatus, fingerprint identifying..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fingerprint registering apparatus, fingerprint identifying... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2473076