JP4245093B2 - Fingerprint verification device and processing method thereof - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an apparatus for processing fingerprint registration and fingerprint verification by processing an input fingerprint image and a processing method thereof.
[0002]
[Prior art]
In order to prohibit entry of persons other than those concerned into the specific area in the building, a door that is opened and closed by an electric lock is provided at the passage or entrance to the specific area, and access to the specific area is permitted in the vicinity of this door In some cases, a fingerprint verification device is installed to determine whether or not.
FIG. 15 is an overall configuration diagram showing a conventional fingerprint collation apparatus disclosed in, for example, Japanese Patent Laid-Open No. 3-127192. Since the details of this apparatus will be described later, an outline of the operation will be described here.
[0003]
When a passerby puts a finger on the fingerprint sensor 1, the fingerprint sensor 1 reads a fingerprint image, which is stored in the image memory 2. Then, in order to obtain the feature points of the fingerprint image, the image feature / feature extraction means 3 extracts the feature points of the fingerprint. In order to extract the feature points, for example, the template 14 shown in FIG. 3 is extracted by scanning the fingerprint image 1a pixel by pixel as shown in FIG.
[0004]
Here, if the template 14 is scanned for each pixel, the image processing time becomes enormous. Therefore, the ridge pitch P of the fingerprint shown in FIG. Actually, the scanning lines L1 to L3 shown in FIG. 6 are run, and the average value of the ridge pitches P determined for each of the scanning lines L1 to L3 is measured. The scanning pitch at the time of extraction is determined. FIG. 7 shows the relationship between the ridge pitch P and the image processing scanning pitch.
[0005]
Further, the determined scanning pitch P is input to the image processing / feature extraction means 3. If the ridge pitch P is large according to the ridge pitch P of the fingerprint, the scanning pitch for image processing is coarse and the ridge pitch is small. Then, the scanning pitch of the image processing is made fine, the image processing is performed at the optimum scanning pitch, and the feature point of the fingerprint is extracted.
Then, at the time of fingerprint registration, the fingerprint information from the image processing / feature extraction unit 3 and the scanning pitch are registered by the fingerprint registration unit 4.
[0006]
At the time of fingerprint collation, image processing and feature point extraction are performed by setting a scanning pitch via the scanning pitch setting means 7 based on the ridge pitch P measured by the ridge pitch measuring means 6. The fingerprint collation means 5 collates the fingerprint information at the time of collation with the registered fingerprint information.
[0007]
[Problems to be solved by the invention]
In the conventional fingerprint collation apparatus as described above, since the scanning pitch is set by the average ridge pitch from the ridge pitch measuring means 6, fingerprint image processing is performed at the optimum scanning pitch, but there is no fingerprint image. There is a problem that high-speed processing cannot be expected because unnecessary processing is executed even for a blank portion.
[0008]
In addition, since the optimum scanning pitch is determined only by the ridge pitch P, if there is a fluctuation in the ridge pitch P due to noise or unstable finger placement, image processing is executed at a non-optimal scanning pitch. On the contrary, there is a problem that misrecognition increases. Further, in a collation method (hereinafter referred to as search collation) that enables collation without inputting a complicated password at the time of collation, there is a problem that the pitch information at the time of registration cannot be used as pitch information necessary for collation.
[0009]
The present invention has been made to solve the above-mentioned problems, and provides a fingerprint collation apparatus and a processing method thereof that can speed up image processing and can reliably determine a scanning pitch during collation. Objective.
[0010]
[Means for Solving the Problems]
According to a first aspect of the present invention, there is provided a fingerprint collation apparatus that processes a fingerprint image read by a fingerprint sensor, extracts and registers feature points, and performs fingerprint collation using the registered fingerprint information. A finger area measuring means for measuring a finger area determined from the position of the finger and the area of the finger, a finger area measuring means for measuring the finger area from the finger area measured by the finger area measuring means, A ridge pitch measurement unit that inputs a fingerprint image and measures the ridge pitch of the fingerprint, and image processing of the fingerprint image is performed in a high-resolution and small-area mode only for the finger region measured by the finger region measurement unit. Or mode switching means for selecting whether to perform in a low-resolution and large-area mode. The mode switching means always selects a low-resolution and large-area mode and measures the finger measured by the finger area measuring means. surface The ridge pitch when been ridge pitch and is less than the predetermined value, both measured by the measuring means only, and selects the mode of high resolution and a small area.
[0011]
The fingerprint collation device according to the second invention includes a living body detection device instead of the finger area measuring means, and the living body detection device detects the capacitance of the placed finger and outputs an output proportional to the finger area. To be sent .
[0017]
Further, the fingerprint collation processing method according to the third invention stores a fingerprint image read by a fingerprint sensor, inputs the fingerprint image, measures a finger region determined from the position of the finger and the area of the finger, and measures Measure the finger area from the finger area, input the fingerprint image, measure the ridge pitch of the fingerprint, and select the low resolution and large area mode as the mode for processing the fingerprint image at all times. Only when both the measured finger area and the measured ridge pitch are less than or equal to each predetermined value, the high resolution and small area mode is selected, and the fingerprint image is processed only for the above finger area and the fingerprint feature point Are extracted, feature points of the extracted fingerprint are registered, and fingerprint collation is performed using the registered fingerprint information.
[0019]
DETAILED DESCRIPTION OF THE INVENTION
Embodiment 1
FIGS. 1 to 9 are diagrams showing an embodiment of the first and eighth aspects of the present invention. FIG. 1 is an overall configuration diagram, FIG. 2 is an explanatory diagram of fingerprint feature points, and FIG. FIG. 4 is an explanatory diagram of detection of feature points of a fingerprint, FIG. 5 is an explanatory diagram of a ridge pitch of the fingerprint, FIG. 6 is an explanatory diagram of a measurement position of the ridge pitch of the fingerprint, and FIG. FIG. 8 is an explanatory diagram of image processing of a child finger, FIG. 9 is an explanatory diagram of image processing of an adult finger, and the same reference numerals denote the same parts in the drawings (another embodiment). The same).
[0020]
In FIG. 1, 1 is a fingerprint sensor for reading a fingerprint image 1a of a finger placed on a fingerprint reader (not shown), 2 is an image memory for storing the fingerprint image 1a from the fingerprint sensor 1, and 3 is on the image memory 2 Image processing / feature extraction means for extracting fingerprint feature points by image processing of fingerprint images, 4 is a fingerprint registration means for registering fingerprint information from the image processing / feature extraction means 3, and 5 is registration from the fingerprint registration means 4. The fingerprint collating unit compares the fingerprint information with the fingerprint information from the image processing / feature extracting unit 3 at the time of collation and performs collation processing.
[0021]
6 is a ridge pitch measuring means for processing the fingerprint image on the image memory 2 to measure the ridge pitch of the fingerprint, and 7 is an image processing / feature extracting means based on the average ridge pitch from the ridge pitch measuring means 6. A scanning pitch setting means 8 for setting a scanning pitch of 3; a finger area determined by a finger placement position and a finger area is measured from a fingerprint image on the image memory 2; This is finger area measurement means for giving a designation signal.
[0022]
Next, the operation of this embodiment will be described.
When a passerby puts a finger on the fingerprint sensor 1, the fingerprint sensor 1 reads the fingerprint image 1a, which is input to the image memory 2 and stored for one screen. In order to obtain the feature points of the fingerprint image on the image memory 2, the image processing / feature extraction means 3 removes noise and extracts the feature points of the fingerprint.
[0023]
FIG. 2 shows a representative example of fingerprint feature points. Known feature points of the fingerprint include a branching point 12 where the fingerprint ridge 11 branches and an end point 13 where the fingerprint ridge 11 is interrupted. In order to extract the feature points 12 and 13, for example, the template 14 shown in FIG. 3 is extracted by scanning the fingerprint image 1a pixel by pixel as shown in FIG.
Here, if the template 14 is scanned for each pixel, the image processing time becomes enormous. Therefore, the ridge pitch P of the fingerprint shown in FIG. W is the width of the fingerprint ridge.
[0024]
Actually, as shown in FIG. 6, the scanning lines L1 to L3 are run on the fingerprint image 1a, the average value of the ridge pitches P obtained for each of these scanning lines L1 to L3 is measured, and the scanning pitch setting means 7 determines the scanning pitch for image processing and feature point extraction. FIG. 7 shows the relationship between the ridge pitch P and the image processing scanning pitch. That is, the ridge pitch P is distributed in the range of 100 to 600 μm. For example, if the ridge pitch P is 200 μm, it can be determined that the optimum image processing scanning pitch is 50 μm.
[0025]
Further, the determined scanning pitch is input to the image processing / feature extraction means 3, and if the ridge pitch P is large according to the ridge pitch P of the fingerprint, the scanning pitch for image processing is coarse and the ridge pitch P is small. For example, the scanning pitch of the image processing is made fine, the image processing is performed at the optimum scanning pitch, and the feature points of the fingerprint are extracted.
[0026]
Here, even if the thinning scan pitch for image processing is specified according to only the ridge pitch P of the fingerprint, if the ridge pitch P of a child finger or the like is fine and the area is small, it is effective for speeding up the image processing. Not. That is, in the conventional apparatus, since the processing pitch is simply thinned out, image processing is performed over the entire area. For this reason, with a child's finger or the like, image processing is executed even for an area without a fingerprint, so that it is not possible to increase the processing speed.
[0027]
In this embodiment, for a fingerprint having a small area and a small ridge pitch P, image processing is performed with a high resolution and a small area, thereby realizing a significant increase in speed. In an actual processing circuit, the resolution of the image sensor does not change continuously. For example, if the ridge pitch P is 200 μm, the optimum maximum scanning pitch is 50 μm. In this case, the optimum scanning pitch is 25 μm with respect to the ridge pitch of 100 μm.
[0028]
Therefore, for example, if the resolution of the image sensor and the image memory 2 is 640 × 480 pixels, it corresponds to a child or a person with a fine fingerprint cutting out a partial image on 640 × 480 pixels, and the ridge pitch P is 200 μm. If it is above, it corresponds to performing image processing after reducing the whole. In other words, with a resolution change of about two steps, in reality, most people can be handled. For this reason, in this embodiment, a finger area measuring means 8 is provided.
[0029]
Next, the operation of the finger area measuring means 8 will be described with reference to FIGS. FIG. 8 shows a processing example of a child finger, and FIG. 9 shows a processing example of an adult finger.
For example, assuming that the image memory 2 has a memory amount of 640 × 480 pixels, and there is a child's fingerprint in the fingerprint image 1a input to the image memory 2 as shown in FIG. Although it is necessary to scan, the image processing area 1b may be, for example, 320 × 240 pixels. Accordingly, the scanning pitch of the image processing is set for each pixel, and the image processing area 1b designates 320 × 240 pixels.
[0030]
On the other hand, when there is an adult finger in the fingerprint image 1 a, the image processing area 1 c reduced in size is sent to the image processing / feature extraction means 3. In this case, as shown in FIG. 9, the resolution of the image processing is sufficient for every two pixels, and instead, the image processing area 1c is 640 × 480 pixels. In these two examples, image processing has only to be performed 320 × 240 times, and in the case of a child fingerprint having a small finger area and a small ridge pitch P compared to the conventional case where only the scanning pitch is variable, Compared to the above, the processing time is four times faster.
[0031]
At the time of fingerprint registration, the fingerprint information from the image processing / feature extraction unit 3 and the scanning pitch are registered by the fingerprint registration unit 4. At the time of fingerprint collation, image processing and feature point extraction are performed by setting the scanning pitch via the scanning pitch setting means 7 based on the ridge pitch P measured by the ridge pitch measuring means 6 (measured at the time of search collation). . The fingerprint collation means 5 collates the fingerprint information at the time of collation with the registered information.
[0032]
In this way, not only the scanning pitch of the image processing is optimized by the ridge pitch P of the fingerprint, but also the finger region of the fingerprint is detected, and image processing is not performed on unnecessary portions, so image processing is performed. Time can be shortened.
[0033]
Embodiment 2
FIG. 10 is an overall configuration diagram showing one embodiment of the second, third and ninth inventions of the present invention. 2 to 9 are also used in the second embodiment.
In the figure, 21 is a finger area measuring means for measuring the finger area from the area position signal 8a which is the output of the finger area measuring means 8, 22 is a high resolution image of the fingerprint image on the image memory 2 from the finger area and ridge pitch P. It is a mode switching means for outputting an image processing mode signal 22a for determining whether processing is performed with a small area or processing with a low resolution and a large area.
[0034]
The region position signal 8 a is also output to the image processing / feature extraction unit 3, and the image processing mode signal 22 a is output to the image processing / feature extraction unit 3 and the fingerprint registration unit 4. The output of the fingerprint registration unit 4 is also output to the image processing / feature extraction unit 3. The rest is the same as FIG.
[0035]
Next, the operation of parts different from FIG. 1 will be described.
After the fingerprint image 1a is input to the image memory 2, when the region position signal 8a is output from the finger region measuring unit 8, the finger area measuring unit 21 measures the finger area. Depending on the measured finger area and the ridge pitch P measured by the ridge pitch measurement means 6, the mode switching means 22 performs image processing and feature point extraction processing in a high-resolution and small-area mode or a low-resolution mode. Choose whether to run in large area mode.
[0036]
The selection logic here is processing with high resolution and a small area only when the fingerprint area is small and the ridge pitch P of the fingerprint is equal to or smaller than a certain value in consideration of the use of a child as shown in FIG. Select. Further, the image processing / feature extraction unit 3 selects and executes one of high resolution, small area, low resolution, and large area by the image processing mode signal 22a from the mode switching unit 22.
[0037]
Further, at the time of fingerprint registration, the fingerprint registration means 4 stores only the mode signal 22a together with the fingerprint information, and at the time of fingerprint verification, the fingerprint verification means 5 stores the mode signal 22a stored at the time of registration by an operation such as input of a personal identification number. And fingerprint information is obtained, image processing is performed by the mode signal 22a, and collation processing is performed by using the registered fingerprint information and fingerprint information at the time of collation.
[0038]
As described above, in the conventional apparatus, the scanning pitch of the image processing is continuously variable only by the ridge pitch P of the fingerprint. However, since the resolution of about two steps is sufficient in actual use, the image processing is performed at a high resolution. In addition, a mode with a small area or a mode with a low resolution and a large area is provided, and the mode is selected according to the ridge pitch P and the finger area while paying attention to the finger area. It is possible to reduce erroneous determination of resolution selection.
[0039]
Embodiment 3
FIG. 11 is an overall configuration diagram showing an embodiment of the fourth aspect of the invention. 2 to 9 are also used in the third embodiment.
In the figure, reference numeral 23 denotes a living body detection device that detects a living body (finger) by electrostatic capacity, and is connected to the mode switching means 22. The rest is the same as FIG. However, the finger area measuring means 21 is not used.
[0040]
Next, the operation of parts different from FIG. 10 will be described.
The living body detection device 23 detects a living body (finger) by electrostatic capacity and sends an output proportional to the area where the finger is placed to the mode switching means 22. Thus, the mode switching means 22 performs image processing and feature point extraction processing in a high-resolution and small-area mode based on the finger area from the living body detection device 23 and the ridge pitch P from the ridge pitch measurement means 6. Alternatively, it is selected whether to execute in a low resolution and large area mode. Other than the above, the configuration is the same as that of FIG.
[0041]
In this way, the finger is detected based on the capacitance and the finger area is output, and the above mode is selected based on the ridge pitch P and the finger area. Can be reduced.
[0042]
Embodiment 4
FIG. 12 is an overall configuration diagram showing an embodiment of the fifth invention of the present invention. 2 to 4, 8, and 9 are also used in the fourth embodiment.
In the figure, reference numeral 24 denotes a switch or the like, and a mode input device that outputs an image processing mode signal 24a for selecting whether the fingerprint image 1a is executed in a high-resolution and small-area mode or a low-resolution and large-area mode. It is. The rest is the same as FIG. However, the ridge pitch measuring means 6 and the finger area measuring means 21 are not used.
[0043]
In the conventional device, the image processing mode signal and ridge pitch information are stored together with the fingerprint information at the time of registration, and the mode signal and ridge pitch are set at the time of collation input operation, etc. at the time of collation, image processing and collation Processing is realized. However, recently, search collation is being adopted in which fingerprint data is retrieved and collated by simply placing a finger without inputting a password at the time of collation. However, in this search and collation, since the setting information at the time of registration as described above cannot be used, it is necessary to obtain the ridge pitch information of the fingerprint and other information independently at the time of collation.
[0044]
At this time, if the ridge pitch information and the mode information do not match at the time of registration and at the time of matching, there is a problem that the matching rate is extremely lowered. The fourth embodiment intends to input the mode state independently in order to deal with the above-mentioned problem.
Next, the operation of parts different from FIG. 10 will be described.
When inputting the mode state, the mode input device 24 is set manually.
[0045]
For example, when a child uses it, the mode signal 24a is always set to a high-resolution and small-area mode by turning on the switch. For adults, the mode is switched to a low-resolution and large-area mode by turning off the switch. Set. Other than the above, the configuration is the same as that of FIG.
Since the image processing mode is directly input in this manner, the processing mode is surely selected, and erroneous determination of the image processing mode setting at the time of search collation can be eliminated.
[0046]
Embodiment 5
FIG. 13 is an overall configuration diagram showing an embodiment of the sixth aspect of the invention. 2 to 4, 8, and 9 are commonly used in the fifth embodiment.
In the figure, reference numeral 25 denotes a height measuring device for measuring a height by blocking light beams arranged at a certain height when passing, and is connected to the mode switching means 22. The rest is the same as FIG. However, the mode signal 24a from the mode input device 24 to the fingerprint registration device 4 and the signal from the fingerprint registration unit 4 to the image processing / feature extraction unit 3 are not used.
[0047]
This embodiment automates the mode switching of the fourth embodiment by a more reliable element, and there is usually a correlation between height, fingerprint ridge pitch, and finger size in adults and children. It is what paid attention to.
[0048]
Next, the operation of this embodiment will be described.
The height measuring device 25 measures the height of the passerby and sends it to the mode switching means 22. The mode switching means 22 sets the image processing mode signal 22a to a high-resolution and small-area mode if the height is equal to or less than a predetermined threshold, and low-resolution and large-area if the height exceeds the predetermined threshold. Set the mode. Since other than the above is similar to FIG.
[0049]
In this way, since the image processing mode is input based on the height, the processing mode is surely selected, and it is possible to eliminate erroneous determination of the image processing mode setting at the time of search collation.
[0050]
Embodiment 6
FIG. 14 is an overall configuration diagram showing an embodiment of the seventh invention of the present invention. 2 to 4, 8, and 9 are also used in the sixth embodiment.
In the figure, reference numeral 26 denotes a body weight measuring device for measuring body weight when passing, and is connected to the mode switching means 22. The rest is the same as FIG.
In this embodiment, the mode switching of the fourth embodiment is automated by a more reliable element, and there is a correlation between the weight, the ridge pitch of the fingerprint, and the finger size in normal adults and children. It is what paid attention to.
[0051]
Next, the operation of this embodiment will be described.
The weight measuring device 26 measures the weight of the passerby and sends it to the mode switching means 22. The mode switching means 22 sets the image processing mode signal 22a to a high-resolution and small-area mode if the weight is equal to or less than a predetermined threshold, and low-resolution and large-area if the weight exceeds the predetermined threshold. Set the mode. Since other than the above is similar to FIG.
[0052]
Thus, since the image processing mode is input by the body weight, the processing mode is surely selected, and it is possible to eliminate erroneous determination of the image processing mode setting at the time of search collation.
[0053]
【The invention's effect】
As described above, in the first invention of the present invention, a fingerprint image is input and a finger region is measured, and image processing of the fingerprint image is performed only in the finger region in a high resolution and small area mode, or When the low resolution and large area mode is selected , the low resolution and large area mode is selected at all times, and the measured finger area and ridge pitch are both below the specified values. However, since the high-resolution and small-area mode is selected, it is possible to increase the speed of image processing and to reduce erroneous determination of resolution selection.
[0054]
In the second invention, since the mode is switched according to the finger capacitance and the ridge pitch, it is possible to increase the speed of image processing and reduce erroneous determination of resolution selection.
[0056]
In the third invention, image storage, the finger domain measurements, the finger area measurement, ridge pitch measurement, mode switching, image processing and feature extraction, since e Bei each step of the fingerprint registration and fingerprint matching, high-speed image processing And erroneous determination of resolution selection can be reduced.
[Brief description of the drawings]
FIG. 1 is an overall configuration diagram showing Embodiment 1 of the present invention.
FIG. 2 is an explanatory view of feature points of a fingerprint showing Embodiment 1 of the present invention.
FIG. 3 is a plan view of a fingerprint feature point detection template according to Embodiment 1 of the present invention.
FIG. 4 is an explanatory diagram of fingerprint feature point detection according to Embodiment 1 of the present invention;
FIG. 5 is an explanatory diagram of fingerprint ridge pitches according to the first embodiment of the present invention.
FIG. 6 is an explanatory diagram of a ridge pitch measurement position of a fingerprint showing Embodiment 1 of the present invention.
FIG. 7 is a relationship diagram between fingerprint ridge pitches and image processing scanning pitches according to the first embodiment of the present invention;
FIG. 8 is an image processing explanatory diagram of a child finger showing Embodiment 1 of the present invention.
FIG. 9 is an explanatory diagram of adult finger image processing according to the first embodiment of the present invention.
FIG. 10 is an overall configuration diagram showing Embodiment 2 of the present invention.
FIG. 11 is an overall configuration diagram showing Embodiment 3 of the present invention.
FIG. 12 is an overall configuration diagram showing Embodiment 4 of the present invention.
FIG. 13 is an overall configuration diagram showing Embodiment 5 of the present invention.
FIG. 14 is an overall configuration diagram showing Embodiment 6 of the present invention.
FIG. 15 is an overall configuration diagram showing a conventional fingerprint collation apparatus.
[Explanation of symbols]
1 fingerprint sensor, 1a to 1c fingerprint image, 2 image memory, 3 image processing / feature extraction means, 4 fingerprint registration means, 5 fingerprint collation means, 6 ridge pitch measurement means, 7 scanning pitch setting means, 8 finger area measurement means , 11 Fingerprint ridges, 12, 13 fingerprint feature points, 21 Finger area measuring means, 22 Mode switching means, 23 Living body detection device. 24 mode input device, 25 height measuring device, 26 weight measuring device

Claims (3)

In a device that processes a fingerprint image read by a fingerprint sensor, extracts and registers a feature point, and performs fingerprint verification using this registered fingerprint information,
A finger region measuring means for inputting the fingerprint image and measuring a finger region determined from the position of the finger and the area of the finger;
Finger area measuring means for measuring the finger area from the finger area measured by the finger area measuring means;
A ridge pitch measuring means for inputting the fingerprint image and measuring a ridge pitch of the fingerprint;
Mode switching means for selecting whether the image processing of the fingerprint image is performed in the high resolution and small area mode or the low resolution and large area mode only for the finger area measured by the finger area measuring means. It equipped with a door,
The mode switching means always selects a mode of low resolution and large area, and the finger area measured by the finger area measuring means and the ridge pitch measured by the ridge pitch measuring means are both predetermined values. A fingerprint collation device that selects a high-resolution and small-area mode only in the following cases . Comprising a biological detection device in place of the finger area measurement means, the biological detection device detects the electrostatic capacitance of the placed finger, according to claim 1, wherein sending an output proportional to the finger area Fingerprint verification device. An image storage step for storing a fingerprint image read by the fingerprint sensor;
A finger region measurement step of measuring the finger region determined from the position where the finger is placed and the area of the finger by inputting the fingerprint image;
A finger area measuring step of measuring a finger area from the measured finger region;
A ridge pitch measurement step for measuring the ridge pitch of a fingerprint by inputting the fingerprint image;
As a mode for image processing of fingerprint images, a low resolution and large area mode is always selected, and only when the measured finger area and the measured ridge pitch are less than or equal to each predetermined value, the high resolution is selected. And a mode switching step for selecting a mode of a small area ;
An image processing / feature extraction step for inputting the stored fingerprint image, processing the fingerprint image only for the measured finger region, and extracting a feature point of the fingerprint;
A fingerprint registration step of registering the extracted feature points of the fingerprint;
A fingerprint collation processing method comprising: a fingerprint collation step of collating a fingerprint using the registered fingerprint information.

Images