JP2875053B2 - How to update registered fingerprint feature points - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for updating registered fingerprint feature points, and more particularly, to a method for updating registered fingerprint feature points used in a personal identification device using a fingerprint such as an information security system.

[0002] In recent years, with the introduction of computers into a wide range of social systems, concerned parties have been interested in system security. ID cards and passwords that have been used as a means of personal identification when entering a computer room or using a terminal have raised many questions in terms of security.

[0003] On the other hand, fingerprints are considered to be the most effective means of personal identification because they have two major characteristics, "everyone is unidentified" and "lifelong invariant". R & D is underway.

[0004]

2. Description of the Related Art FIG. 10 is a block diagram of a conventional personal identification device, in which 1 is a fingerprint sensor, 2 is a binarization circuit, and 3 is 2
4 is a registration unit, 5 is a feature information storage unit, 6 is a matching unit, 7 is a main displacement amount changing unit, 8 is a sub displacement amount changing unit, 9 is a memory address specifying unit, 10 is a pixel value comparing unit. , 11 denotes a pass window evaluation unit, and 12 denotes a feature extraction unit.

Conventionally, for example, a personal identification device using the moving window method includes a fingerprint sensor 1, a binarization circuit 2, a binarization memory 3, a registration unit 4, a feature information storage unit 5, a collation unit, as shown in FIG. And the like.

[0006] The registration unit 4 includes a feature extraction unit 1.
2 and the collation unit 6 includes a main displacement amount changing unit 7,
A sub displacement amount changing unit 8, a memory address specifying unit 9, a pixel value comparing unit 10, and a pass window number evaluating unit 11 are provided.

[0007] In the above device, when registering a fingerprint, the following is performed. First, a fingerprint image is input from the fingerprint sensor 1. This input image is converted into a 2
After performing the binarization process, it is stored in the binarization memory 3.

After that, the feature extracting unit 12 extracts a feature point of the fingerprint from the fingerprint image (binary image) in the binarized memory 3 and registers it in the feature information storage unit 5 as personal feature information.

At the time of this feature point extraction, a fingerprint image near a feature point such as an end point or a branch point of the fingerprint image is cut out as a “window”, and for each “window”, the position coordinates of the “window” and the “window” are extracted.
Are registered as dictionary information of personal characteristic information.

The process at the time of fingerprint collation is as follows.
First, a fingerprint image is input by the fingerprint sensor 1 and binarized by the binarization circuit 2.
Store the digitized fingerprint image.

After that, the collating unit 6 verifies the identity by collating the input image (binary image of the fingerprint) in the binarization memory 3 with the dictionary information in the feature information storage unit 5.

In this case, first, pattern matching is performed between the positioning window and the input image to correct the difference between the input position at the time of registration and the input position at the time of verification. Next, according to the displacement amount of the alignment window, the surrounding matching windows are translated.

These processes are executed by the main displacement amount changing unit 7 and the memory address specifying unit 9
This is performed by sending an address to the feature information storage unit 5.

By sending this address, the binary memory 3
And the window image information (dictionary information) from the feature information storage unit 5 into the pixel value comparison unit 10 to compare the pixels.

However, in this comparison processing, the coincidence between the two pixels is not very good (because the finger is stretched and shrunk). For this reason, the sub-displacement amount changing unit 8 moves each collation window independently and slightly, and performs the same pixel comparison as above.

The pass / fail information of the "window", which is the information of the result of the pixel comparison, is taken into the pass window evaluation section 11 and the pass / fail evaluation is performed. By this evaluation, more than a predetermined number of "windows"
Is passed, an identity confirmation signal (a signal indicating that the identity has been confirmed) is output.

[0017]

The above-mentioned conventional apparatus has the following problems.

(1) In a conventional personal identification device, fingerprint characteristics are registered in advance as a dictionary, and at the time of collation, an input image is compared with characteristic dictionary information to perform personal identification.

However, in general, at the time of registration, the user is often unfamiliar with the device, and the position where the finger is placed and the pressure are usually not constant.

Further, it is often difficult to input a fingerprint image having the same quality as that at the time of registration at the time of collation due to a change in the work content of the user after registration or a seasonal change in skin condition. .

(2) On the other hand, there is also a method in which a fingerprint image is input a plurality of times at the time of registration, a dictionary is created for each image, and a feature point commonly selected for each image is preferentially adopted as the dictionary. Proposed.

According to this method, it is possible to create a dictionary having a margin with respect to the difference in the pressed position and the pressed pressure of the finger, but it is difficult to cope with the aging of the fingerprint.

An object of the present invention is to solve such a conventional problem and to sufficiently cope with a change in pressure at the time of inputting a fingerprint and a change with time of the fingerprint.

[0024]

FIG. 1 is a diagram showing the principle of the present invention, in which the same reference numerals as those in FIG. 10 denote the same parts. Reference numeral 13 denotes an update dictionary temporary storage unit, and reference numeral 14 denotes an update necessity determination unit.

The present invention has the following configuration in order to solve the above problems.

(1) A feature information storage unit 5 for registering personal feature information of a fingerprint in advance as a dictionary, and a feature point is extracted from a fingerprint image, and the information of the fingerprint feature point is used as feature information as personal feature information. A registered personal identification device including a registration unit 4 for registering in the information storage unit 5 and a matching unit 6 for matching the fingerprint features registered in the feature information storage unit 5 with the input fingerprint features. In the fingerprint feature point update method, when fingerprint matching is successful, the remaining feature points are also compared.
What determines the acceptability of the feature points of the dictionary, based on the results, counts the number of bad feature points, the count value is, prefix
The dictionary is updated when the threshold for document update is exceeded, and the dictionary is updated.
It is determined not to update the dictionary if the new threshold is not exceeded.
Perform dictionary update necessity determination of the constant which, if required a dictionary updating from the input fingerprint image during the verification, the newly selected feature points, the registered defective feature points as a dictionary, said <br /> Replaced with the newly selected feature point .

[0027]

(2) When determining the necessity of updating the dictionary, the pass rate is calculated for each feature point based on the past pass rate, the number of times of use, and the result of the current comparison, and the pass rate is low and the pass rate is low. When the number of feature points obtained exceeds the threshold value, it is determined to update the dictionary.

(3) In the configuration (2) , when the dictionary update necessity determination is performed, a lower limit value is set for the number of times the dictionary is used,
When the number of uses is less than the lower limit, it is determined that the dictionary is not updated even when the number of defective feature points having a low pass rate exceeds the threshold value.

(4) When the dictionary update necessity determination is performed, a matching error is calculated for each feature point, and if the number of feature points determined as bad due to a large matching error exceeds a threshold, the dictionary is updated. Judge to update.

(5) When performing dictionary update necessity determination, an average collation error is calculated for each feature point, and when the average collation error is large and the number of bad feature points exceeds a threshold, Is determined to update the dictionary.

(6) In the configuration (5) , when determining whether or not to update the dictionary, a lower limit value is set for the number of times the dictionary is used, and if the number of times of use is less than the lower limit, the average collation is performed. Even when the number of defective feature points having a large error exceeds the threshold value, it is determined not to update the dictionary.

[0033]

The operation of the present invention based on the above configuration will be described with reference to FIG.

(1) At the time of fingerprint collation, the collation unit 6
Dictionary (personal characteristic information) registered in the characteristic information storage unit 5
And the feature point of the input fingerprint stored in the binarization memory 3.

At the time of this collation, a pass / fail judgment is made for each feature point, and the result is stored in an internal memory. If the identity can be confirmed by this collation, the remaining feature points are also collated (identification is determined without collation for all feature points), and the pass / fail result is stored in the internal memory. Keep it.

Thereafter, the update necessity determining unit 14 determines
The pass / fail data is verified to determine whether dictionary update is necessary.
In this process, first, the number of defective feature points (rejected feature points) is counted.

If the number of defective feature points exceeds the dictionary update threshold, it is determined that dictionary update is required.

When updating the dictionary, a new feature point is selected from the input image at the time of collation stored in the binarization memory 3 and is temporarily stored in the update dictionary storage unit 13.

Thereafter, the dictionary registered in the feature information storage unit 5 is rewritten with the newly selected feature point.

In this way, the dictionary is updated as the user becomes more proficient in the personal identification device.
It is possible to sufficiently cope with the temporal change of the fingerprint.

Further, when the dictionary update necessity determination is performed, the above configurations (2) to (7) may be employed. Also in these cases, as described above, it is possible to sufficiently cope with a change in finger pressing and a change with time in fingerprints, and it is possible to always perform stable identity verification.

[0042]

Embodiments of the present invention will be described below with reference to the drawings.

(Explanation of First Embodiment) FIGS. 2 and 3 show a first embodiment of the present invention. FIG. 2 is a block diagram of the personal identification device, and FIG. 3 is a processing flowchart.

In the drawings, the same reference numerals as those in FIGS. 1 and 10 denote the same components. Reference numeral 15 denotes a pass window evaluation unit, and reference numeral 16 denotes an update threshold value setting unit.

The first embodiment is an example in which the present invention is applied to an identity verification apparatus that performs fingerprint collation by the moving window method. FIG. 2A shows a block diagram of the apparatus, and shows the data structure of personal characteristic information. 2 (B).

The personal identification device of this embodiment has an update dictionary temporary storage unit 13 in the registration unit 4 of the conventional device shown in FIG.
And an update necessity determination unit 14, and a collation unit 6.
In addition, a pass window evaluation unit 15 and an update threshold setting unit 16 are added to FIG.

The personal characteristic information registered as a dictionary in the characteristic information storage unit 5 is composed of window position information (X, Y) and window image information.

The update dictionary temporary storage unit 13 temporarily stores information on a “window” near a feature point newly selected from an input fingerprint image in order to update the dictionary.

The pass window number evaluation unit 15 includes a pixel value
The pass / fail information of the "window" output from is input and the number of passing windows is evaluated.

The update threshold value setting unit 16 sends an update threshold value for judging the necessity of updating the dictionary to the pass window number evaluation unit 15,
The threshold value is set.

The update necessity judging section 14 judges whether or not the dictionary needs to be updated based on the evaluation result of the passing window evaluation section 15 and performs a dictionary update process at the time of updating.

In the dictionary update processing, the update necessity determining unit 14 determines whether the update is necessary from the update dictionary temporary storage 13
And updates the dictionary by rewriting the dictionary registered in the feature information storage unit 5.

In this update, the dictionary is updated only for the defective “window” to be a new dictionary.

Hereinafter, the processing of the first embodiment will be described with reference to the processing flowchart of FIG. Each processing number in FIG. 3 is shown in parentheses.

In the processing of FIG. 3, the processing of S1 to S4 is the same as the conventional fingerprint collation processing. However, the processing of S2 is added for the next processing. This will be specifically described below.

When performing fingerprint matching by the moving window method, the pixel value comparison unit 10 extracts a “window” cut out from the vicinity of a feature point of the input fingerprint and the feature information storage unit 5.
Next, a comparison is made between the pixel value and the "window" registered as personal characteristic information (dictionary) (S1).

When the comparison process is completed, pass / fail information of “window” is output.
1 and 15 (S2).

The above processing is performed for all "windows" (S4), and when the number of passing windows exceeds a certain number (S4).
3), the personal identification processing ends. This identity verification process
As in the conventional case, the pass window number evaluation unit 11 performs the process. In addition,
If the number of passed "windows" does not reach the number required for identity verification even after collation for all "windows", rejection is performed.

When the fingerprint collation succeeds (when the identity is successfully confirmed), collation is also performed for the remaining “windows” (S5), and the pass / fail information is passed to the pass window evaluation unit 1.
5 (S6).

When the collation of the remaining "windows" is completed, the pass / fail evaluation unit 15 verifies the pass / fail data (S7). In this verification, the pass / fail data for each “window” stored in advance is checked, and if any “window” has failed (S8), the number is counted (S9).

After this process has been performed for all "windows" (S10), the update necessity determining unit 14 determines whether or not the dictionary needs to be updated (S11).

In this case, the number-of-accepted-windows evaluation section 15 compares the threshold value set by the update threshold value setting section 16 with the count value of the rejected “window”, and the count value exceeds the threshold value. In this case, the output signal to that effect is determined by the determination unit
4

The update necessity judging section 14 receives the output signal and judges the necessity of update. That is, when the count value of the rejected “window” exceeds the threshold value, it is determined to update the dictionary.

When it is determined that the dictionary needs to be updated (S11), a new fingerprint image is used from the fingerprint image used at the time of collation.
Select "window" and update the dictionary.

That is, a thinning process is performed using the input fingerprint image (the fingerprint image used at the time of matching) stored in the binarization memory 3 (S12), and feature points are extracted (S13).

Thereafter, a “window” is cut out from the vicinity of the feature point (S 14), and this is temporarily stored in the update dictionary temporary storage unit 13.

Next, the dictionary which has been registered in the feature information storage unit 5 is updated by the update necessity determination unit 14 using the “window” information stored in the update dictionary temporary storage unit 13. I do. In this dictionary update, only the failed “window” is updated.

(Explanation of the Second Embodiment) FIG. 4 is a block diagram of the personal identification device in the second embodiment, and FIG. 5 is a processing flowchart.

In the figure, the same reference numerals as those in FIG. 2 denote the same parts.
Reference numeral 17 denotes a pass rate calculation unit.

The second embodiment is an example in which the present invention is applied to a personal identification device that performs fingerprint collation by the moving window method. FIG. 4A is a block diagram of the personal identification device, and FIG. 3 shows a data structure of feature information.

The personal identification device of this embodiment is provided in the registration unit 4 of the conventional device shown in FIG.
3 and an update necessity determination unit 14, and a pass rate calculation unit 17 is added to the collation unit 6.

The personal characteristic information includes window position information (X, Y), pass rate, number of times of use, window image information (binary)
It consisted of.

The pass rate calculating section 17 uses the past pass rate and the number of times of use registered in the feature information storage section 5 and the information of the current collation result output from the pixel value comparing section 10. The pass rate is calculated for each feature point.

The other configurations are the same as those shown in FIGS. 2 and 10, and a description thereof will be omitted.

Next, the processing of the second embodiment will be described with reference to the processing flowchart of FIG. Each processing number in FIG. 5 is shown in parentheses.

In the processing of FIG. 5, the processing of S20 to S25 and the processing of S31 to S34 are the same as the processing of S1 to S6 and the processing of S12 to S15 in FIG. However, the storage of the pass / fail result (S21, S25) is performed in the pass rate calculation unit 17.

When the storage of the pass / fail results for all the “windows” is completed (S25), the pass rate is calculated for each “window” using the pass rate history information in the pass rate calculating section 17 (S26). ).

The history information in this case is stored in the feature information storage unit 5.
Use personal characteristic information (pass rate in the past, number of uses).

Next, the calculated pass rate for each “window” is sent to the update necessity judging unit 14, where each “window” needs to be updated based on the pass rate. Or not (pass / fail)
Is determined (S27), and the number of defective “windows” having a low pass rate is counted (S29).

After this process has been performed for all “windows” (S28), the update necessity judging unit 14 judges whether the dictionary needs to be updated (S30).

In this dictionary update necessity determination, it is determined whether the number of defective “windows” having a low pass rate exceeds a prescribed value and whether the number of uses exceeds a predetermined number. .

When the number of “windows” determined to be defective exceeds a prescribed value and the number of times of use (the number of times of use of the dictionary) exceeds a certain number, it is determined to update the dictionary. .

Therefore, even if the number of defective “windows” exceeds the specified value, if the number of uses is less than a certain number,
Do not update the dictionary.

(Explanation of Third Embodiment) FIG. 6 is a block diagram of the personal identification device in the third embodiment, and FIG. 7 is a processing flowchart.

In the figure, the same reference numerals as those in FIG. 4 denote the same parts.
Reference numeral 18 denotes a collation error temporary storage unit.

The personal identification device according to the third embodiment is a device for collating a fingerprint by the moving window method, and its block diagram is shown in FIG. 6A, and the data structure of the personal characteristic information is shown in FIG. ).

The personal identification device of this embodiment has an update dictionary temporary storage unit 13 in the registration unit 4 of the conventional device shown in FIG.
And an update necessity determining unit 14 are provided.
This is provided with a collation error temporary storage unit 18.

As the personal characteristic information, window position information and window image information are used.

The above-described collation error temporary storage unit 18 temporarily stores the collation error (pixel value collation error), which is the collation result in the pixel value comparison unit 10, for each “window”. .

The other components are the same as those shown in FIGS. 2 and 4, and the description is omitted.

Hereinafter, the processing of the third embodiment will be described with reference to the processing flowchart of FIG. Each processing number in FIG. 7 is shown in parentheses.

In the process shown in FIG. 7, the collation errors are stored in the collation error temporary storage unit 18 in the processes of S41 and S45, and in the verification of the pass / fail data in S46, each of the stored “windows” is verified. The collation error is verified, and in the determination of S50, the necessity of dictionary update is determined based on the verification result, and the other processing is the same as the processing of FIG.

More specifically, it is as follows. At the time of collation, each “window”
The pixel values are compared every time, and the matching error (pixel value matching error) at this time is stored in the matching error temporary storage unit 18 (S41).

If the identity can be confirmed, collation is further performed on the remaining "windows", and the collation error is stored in the same manner as described above (S45).

After performing the above-described processing for all “windows”, the update necessity determining unit 14 determines, for each “window”, whether a collation error exceeds a certain threshold (S46).

If the collation error exceeds the threshold (S
47) Assuming that the “window” is a defective “window”, the number of the “window” is counted (S49).

After this processing is performed for all “windows” (S
48) If the number of defective “windows” exceeds a predetermined threshold value, it is determined to update the dictionary (S50).

The processing for updating the dictionary is the same as in the first and second embodiments.

(Explanation of the Fourth Embodiment) FIG. 8 is a block diagram of the personal identification device in the fourth embodiment, and FIG. 9 is a processing flowchart.

In the figure, the same reference numerals as those in FIG. 6 denote the same parts.
Reference numeral 19 denotes an average collation error calculation unit.

The personal identification device according to the fourth embodiment is a device for collating a fingerprint by the moving window method, and its block diagram is shown in FIG. 8A, and the data structure of the personal characteristic information is shown in FIG. ).

The identity verification device of this embodiment is obtained by adding an average collation error calculator 19 to the device shown in FIG. 6, and the other configuration is the same as that of FIG.

The personal characteristic information includes window position information (X,
Y), the average collation error, the number of uses, and the image information of the window.

The above-mentioned average collation error calculating section 19 compares the collation error for each “window” stored in the collation error temporary storage section 18 with the past average collation error stored in the feature information storage section 5 and the The number of times is input, and an average collation error is calculated for each “window”.

Next, the processing of the fourth embodiment will be described with reference to the processing flowchart of FIG. Each processing number in FIG. 9 is shown in parentheses.

In the process of FIG. 9, the average collation error is calculated in S66, and based on the calculation result, it is determined in S67 whether each “window” is a defective “window”. Other processes are the same as those in FIG.

A more specific description is as follows.
In the process of S66, the average collation error for each “window” is calculated. In this case, the average collation error calculator 19 compares the past collation error, the number of times of use (data registered as personal feature information) and the current , A new average collation error is calculated for each “window” using the collation error (data in the collation error temporary storage unit).

In the process of S67, the update necessity determining unit 14 is determined.
In step (1), the average collation error calculated by the average collation error calculation unit 19 is input, and it is determined whether or not each "window" is a defective "window".

If the average collation error exceeds a preset threshold, it is determined that the window is a defective “window”. If the average collation error is less than the threshold, it is determined that the window is not a defective “window”. If there is a defective "window", the number of the "window" is counted (S6).
8).

After this processing is performed for all “windows” (S
69) If the number of defective "windows" exceeds a predetermined value, it is determined that the dictionary needs to be updated (S7).
0). However, if the number of uses is less than a certain number, the dictionary is not updated.

(Other Embodiments) Although the embodiments have been described above, the present invention can be implemented as follows.

(1) The present invention can be applied not only to the personal identification device using the moving window method but also to the personal identification device using another method.

In particular, of the above embodiments, the methods of the first and second embodiments shown in FIGS. 2 to 5 can be applied to an apparatus for performing identity verification by a method other than the moving window method.

(2) For example, as a device for performing identity verification by a method other than the moving window method, a “fingerprint collation system using valley thinning” (“Transactions of the Institute of Electronics, Information and Communication Engineers” D,
Vol, J71-D No.2, PP327-335, see February 1988)
The present invention is also applicable to this system.

[0115]

As described above, the present invention has the following effects.

(1) The dictionary (individual characteristic information) is updated as the user becomes proficient in the personal identification device, so that it is possible to cope with a change in the imprint position of the finger, a change in the press, or a change with time in the fingerprint.

(2) Since an appropriate dictionary can always be registered, fingerprint collation can be stably performed.

(3) Since the number of defective feature points in the dictionary can be reduced, the identity can be quickly confirmed. In addition to the above effects,
The following effects are correspondingly obtained. (4) Claim 1 has the following effects. That is, the registered finger
The more feature points that match between the crest and the input fingerprint, the more
Their fingerprint matching performance is stable. Therefore, the number of defective feature points
The stability of the matching performance of the dictionary by counting
Can be evaluated and the time of dictionary update can be known. Bad
The feature points judged to be points are the state of the finger when the dictionary was created.
Is expected to be bad.
By creating a dictionary, it is updated to excellent feature points. Immediately
That is, an improvement in dictionary collation performance can be expected. Also, input finger
Recreate the dictionary from the crest image and update the existing dictionary
The more you use, the more dictionary of good feature points
Addition can be expected. In addition, the finger that occurs throughout the spring, summer, autumn and winter
It can respond to a change in state. (5) Claim 2 has the following effects. That is, each feature
Since the pass rate is calculated for each point taking into account the past history,
The certainty of good feature points is guaranteed. For example, bad finger condition
When fingerprint matching is performed with
There is a feature point dictionary that does not match at the feature point,
It is possible that it will exceed the dictionary update threshold
You. In such a case, if the past history is not
These excellent feature point dictionaries will also be updated, there
With reference to the past history,
By not determining bad feature points, incorrect dictionary updates
Can be prevented. (6) Claim 3 has the following effects. For example, a dictionary
If the number of uses is small, it can be judged that it is a bad feature point
Erroneous dictionary update occurs because only history is not collected
It is possible. Therefore, the number of times the dictionary is used exceeds the threshold
By updating the dictionary when the
Such an erroneous dictionary update can be prevented. (7) Claim 4 has the following effects. That is, in the dictionary
Whether the included feature points are bad or not
Evaluate based on matching error instead of evaluating based on match or mismatch
By doing so, the certainty of the defective feature point is guaranteed. Toes
Not only match, not match, but how much
It can be understood quantitatively. Set a threshold value for this,
It becomes possible to limit to defective feature points. (8) In claim 5, similar to claim 2, the past history is retained.
This guarantees the certainty of the defective feature point. (9) Claim 6 has the following effects. That is, use times
If the number is less than the lower limit, the average matching error is large and
Dictionary update even if the number of bad feature points exceeds the threshold
Is determined not to be performed.
In addition, erroneous dictionary update can be prevented.

[Brief description of the drawings]

FIG. 1 is a principle diagram of the present invention.

FIG. 2 is a configuration diagram of the personal identification device in the first embodiment.

FIG. 3 is a processing flowchart of the first embodiment.

FIG. 4 is a configuration diagram of a personal identification device in a second embodiment.

FIG. 5 is a processing flowchart of a second embodiment.

FIG. 6 is a configuration diagram of a personal identification device in a third embodiment.

FIG. 7 is a processing flowchart of a third embodiment.

FIG. 8 is a configuration diagram of a personal identification device in a fourth embodiment.

FIG. 9 is a processing flowchart of a fourth embodiment.

FIG. 10 is a configuration diagram of a conventional personal identification device.

[Explanation of symbols]

 Reference Signs List 3 Binarized memory 4 Registration unit 5 Feature information storage unit 6 Collation unit 12 Feature extraction unit 13 Update dictionary temporary storage unit 14 Update necessity determination unit 15 Pass window evaluation unit 16 Update threshold setting unit 17 Pass rate calculation Unit 18 collation error temporary storage unit 19 average collation error calculation unit

Claims (6)

(57) [Claims] 1. A feature information storage unit (5) for registering personal feature information of a fingerprint in advance as a dictionary, extracting feature points from a fingerprint image, and using the information of the fingerprint feature points as personal feature information. A registration unit (4) for registering in the characteristic information storage unit (5); a collation unit (6) for collating the fingerprint characteristics registered in the characteristic information storage unit (5) with the input fingerprint characteristics. In the method of updating the registered fingerprint feature points of the personal identification device equipped with, when the fingerprint matching is successful, the remaining feature points are also matched.
Is performed, and the pass / fail of each feature point in the dictionary is determined. Based on the result, the number of defective feature points is counted, and the total
Updates the dictionary when the numerical value exceeds the dictionary update threshold.
If the dictionary update threshold is not exceeded, the dictionary is updated.
Perform determining dictionary update necessity determination so as not, in a case requiring a dictionary update, from said verification time of the input fingerprint image, newly selected feature points, the registered defective feature points as a dictionary, the A method of updating a registered fingerprint feature point, characterized by replacing with a newly selected feature point. 2. When the dictionary update necessity is determined, based on the past pass rate, the number of times of use, and the result of the current collation,
2. The method according to claim 1, wherein a pass rate is calculated for each feature point, and when the number of feature points determined to be defective due to a low pass rate exceeds a threshold, it is determined to update the dictionary. How to update registered fingerprint feature points. 3. When the dictionary update necessity is determined, a lower limit value is set for the number of times the dictionary is used, and if the number of uses is less than the lower limit, the pass rate is low and the feature is determined to be defective. Even if the number of points exceeds the threshold,
3. The method for updating registered fingerprint feature points according to claim 2, wherein it is determined not to update the dictionary. 4. When a dictionary update necessity determination is performed, a collation error is calculated for each feature point, and when the number of feature points determined to be bad due to a large collation error exceeds a threshold, the dictionary update is performed. 2. The method for updating registered fingerprint feature points according to claim 1, wherein the determination is made to perform the following. 5. When dictionary update necessity determination is performed, an average collation error is calculated for each feature point, and when the average collation error is large and the number of bad feature points exceeds a threshold value, The method for updating registered fingerprint feature points according to claim 1, wherein it is determined that the dictionary is to be updated. 6. When a dictionary update necessity determination is made, a lower limit value is set for the number of times of use of the dictionary, and if the number of times of use is less than the lower limit value, the average matching error is large and is determined to be bad. 6. The method for updating registered fingerprint feature points according to claim 5, wherein it is determined not to update the dictionary even when the number of feature points exceeds a threshold value.