JP2903788B2 - Fingerprint image input device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fingerprint image input device for inputting a fingerprint image for collating a fingerprint.

[0002]

2. Description of the Related Art Heretofore, there has been known a fingerprint collecting apparatus that optically collects a fingerprint image from a subject's finger in order to collate a fingerprint. For example, Japanese Patent Application Laid-Open No. 1-259468 discloses that a fingerprint image is sampled at least twice so that an image taken when the finger is not moving is used for fingerprint collation. It describes that the density difference of the sampled image is calculated for each pixel, and if the density difference is within a certain threshold value, it is determined that the image is a good quality image, and the image is used as an image to be subjected to fingerprint collation. .

In the above publication, as an example of a sampling method, a fingerprint image is sampled for a specific portion, for example, each of four specific lines provided in a dispersed manner.
It is described that the light intensity is detected for each specific line and a fingerprint is collected when a predetermined condition is satisfied for each specific line.

[0004]

However, in the above-described fingerprint collecting apparatus, when a fingerprint is collected, if the finger is not sufficiently pressed against the input surface of the fingerprint input unit, the fingerprint is partially removed. Even if a high-quality fingerprint image was obtained, it was not possible to obtain a fingerprint suitable for performing fingerprint verification.

An object of the present invention is to provide a fingerprint image input device capable of obtaining a fingerprint image suitable for performing fingerprint collation.

[0006]

According to the present invention, there is provided a fingerprint image input device for inputting a fingerprint image for performing fingerprint collation, as shown in FIG.
Having a reading unit that optically reads a fingerprint from the subject's finger,
Fingerprint reading means for obtaining a fingerprint image signal, and a finger on a portion corresponding to each image area of the reading unit based on fingerprint image signals of a plurality of predetermined image areas of the fingerprint image obtained by the fingerprint reading means. Finger detection means for detecting whether or not contact is made, and area determination means for determining, based on an output of the finger detection means, whether or not the area of the fingerprint image is sufficient to perform fingerprint collation. A shape discriminating unit for discriminating whether or not the shape of the fingerprint image is suitable for performing fingerprint collation based on an output of the finger detecting unit; and an area discriminating unit for collating the fingerprint with the area of the fingerprint image. The fingerprint obtained by the fingerprint reading unit is determined in response to the determination that the fingerprint image is sufficient for the fingerprint matching by the shape determining unit. Predetermined image signal Binarization means for converting the binary data using a threshold, the fingerprint image input device comprising the the gist.

[0007]

In the present invention, the fingerprint reading means having the fingerprint reading section optically reads the fingerprint from the finger of the subject to obtain a fingerprint image signal, and the finger detecting means detects the fingerprint image read by the fingerprint reading means. Based on fingerprint image signals of a plurality of image regions set in advance, it is detected whether a finger is in contact with a portion corresponding to each image region of the reading unit,
Based on the output of the finger detecting means, it is determined whether or not the area of the fingerprint image is sufficient to perform fingerprint collation.The shape determining means determines the shape of the fingerprint image based on the output of the finger detecting means. It is determined whether or not the fingerprint image is suitable for fingerprint matching. The area determining means determines that the area of the fingerprint image is sufficient to perform fingerprint matching, and the shape determining means determines the shape of the fingerprint image. Is determined to be suitable for performing fingerprint collation, the binarizing means converts the fingerprint image signal obtained by the fingerprint reading means into binary data using a predetermined threshold.

[0008]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 2 is a schematic block diagram of a fingerprint image input device according to one embodiment of the present invention. 2, the fingerprint image input device includes a fingerprint input unit 1 and a processing unit 2.

The fingerprint input unit 1 as a fingerprint reading means is a device for capturing a fingerprint and outputting a fingerprint image signal, and includes an illumination 11, a prism 12 for fingerprint detection, and an image sensor 13. The illumination 11 irradiates the prism 12 with reference light having a luminance necessary for obtaining optical fingerprint data. The fingertip 3 of the subject is placed on the prism 12, and the reflected light from the prism 12 is photoelectrically converted by the image sensor 13 to create a fingerprint image signal, which is transmitted to the processing unit 2. As the image sensor 13, for example, one having 512 pixels × 512 pixels is used.

The processing section 2 includes an A / D conversion section 21 for converting a fingerprint image signal into a digital signal of multiple gradations, an image memory 22 for storing the converted image data, and an image memory 22.
Calculation / control unit 2 for calculating the difference between the fingerprint image data of the first and the newly input fingerprint image data and controlling each unit
And 3.

Note that all or some of the functions of the processing unit 2 may be included in the fingerprint input unit 1. In addition, in order to perform the image calculation at a high speed, the calculation unit and the control unit of the calculation / control unit 23 may be separated. FIG. 3 is an explanatory diagram showing a memory map of the image memory 22 shown in FIG. 3, the image memory 22 has a storage capacity corresponding to 512 pixels × 512 pixels.

As shown in FIG. 3, predetermined areas of the image memory 22 are used as finger detection monitors m0 to m4. In the finger detection monitors m0 to m4, the fingertip 3 is the prism 12
Area to determine whether or not
In this embodiment, the upper left, upper right, middle, lower left,
The lower right five areas are set. One area is 32 pixels x
It corresponds to 32 pixels.

FIG. 4 is a flowchart showing a process for judging from the outputs of the finger detection monitors m0 to m4 whether the area of the input fingerprint image is sufficient and whether the shape of the fingerprint image is suitable for fingerprint collation. It is. Next, this processing will be described. In step 101, the number n of repeated inputs is initialized, and in step 102, variables j and i are initialized. Subsequently, in step 103, an image of a fingerprint image is input, and the fingerprint image signal is converted into a multi-tone digital signal. The fingerprint image data converted into the digital signal is transferred to the arithmetic and control unit 23.

In step 104, from the fingerprint image data input in step 103, the background image data of the prism 12 which is input before the fingerprint image input and stored in the image memory 22 (the prism image when the finger is not placed) 12 image data) is subjected to a subtraction process. The image created by this subtraction processing is called a gray-scale difference image.

Steps 105 to 109 include:
This is a process of detecting a portion where a finger is placed using the finger detection monitor shown in FIG. First, in step 105, the sum S of the gray values of the finger detection monitor mj (j = 0 at first) is calculated. Next, in step 106, it is determined whether or not the obtained total value S exceeds a threshold U that can be determined to be a finger placed. If the sum S exceeds the threshold U,
In step 107, the variable i is incremented by one. Next, in step 108, all the finger detection monitors m
It is checked whether or not the calculation at j has been completed. If not completed, j is incremented by 1 in step 109 to calculate the sum of the gray values of the next finger detection monitor mj. Return.

In step 106, if the threshold value U for determining that the finger is placed is set to a large value, it is generally necessary to encourage a sweating action to increase the difference in shade, and it takes a long time to input. However, since sweating is sufficient, a stable fingerprint image can be obtained. Conversely, if the threshold U is set small,
Generally, it is possible to input a fingerprint image of a dry finger, but the image quality becomes unstable.

In step 108, if it is confirmed that the calculation of the sum S in all the finger detection monitors mj has been completed, that is, if j = 4, the process proceeds to step 110. In step 110, it is determined whether the total number of i, that is, the total number of finger detection monitors determined to have a finger placed thereon is 3 or more. Here, if i ≧ 3, there is a sufficient image area to perform fingerprint collation, and if i <3, the image area is not enough to perform fingerprint collation.

If i ≧ 3, the routine proceeds to step 111,
It is determined whether the number n of repeated inputs is not 0, and if it is determined that the number n of repeated inputs is not 0, the process proceeds to step 112, where the shape of the fingerprint image detected by the finger detection monitors m0 to m4 is determined. It is determined whether it is appropriate to perform the collation. That is, if it is determined that the shape of the fingerprint image is not appropriate for fingerprint collation even when the total number of finger detection monitors determined to have the finger placed thereon is three or more, step 114 And the number n of repeated inputs is set to 0.

FIG. 5 shows a case where the shape of a fingerprint image detected by the finger detection monitor is appropriate for performing fingerprint collation. FIG. 5A shows a case where a finger is detected by three finger detection monitors. , (B) shows the case where the presence of a finger is detected by four finger detection monitors, and (c) shows the case where the presence of a finger is detected by five finger detection monitors. The case where the matching is not appropriate is a situation other than the detection situation shown in FIG. 5. For example, the finger detection monitor m0, m1, m3 and m4 in FIG. m2, m2, m
4 indicates that a finger is detected, but the finger detection monitors m1 and m3 detect that no finger is present.

The shape of the fingerprint image is determined based on the state of the flag on the finger detection monitor. That is, the state of the flag of the finger detection monitors m0 to m4 when the shape is good as shown in FIG. 5 (for example, “1” when there is a finger, “0” when there is no finger) is stored in the memory in advance. This is performed by determining whether or not the states of the flags of the finger detection monitors m0 to m4 obtained by inputting the fingerprint image match the states of the flags stored in the memory.

If it is determined in step 110 that i <3, the flow advances to step 114 to input a fingerprint image in a stable state, and the number n of repeated inputs is set to zero. . In step 113, the repetition input count n is incremented by one. Step 115
Then, it is determined whether or not the total number of repetition input times n has become equal to the predetermined value N. If they are equal, it is determined that a stable fingerprint image has been input, and the process ends. If the number n of repetitive inputs is not equal to N, the process returns to step 102 in order to execute again from image input.

FIG. 6 is a flowchart showing a process for converting a multi-tone fingerprint image determined to be appropriate for fingerprint collation into a binary image. This process is performed subsequent to the process shown in FIG. Step 201 is performed from the finger detection monitors m0 to m4 determined to have a finger placed thereon as shown in FIG.
This is a process for setting the binarization threshold value determination monitor dm shown in FIG. The binarization threshold determination monitor dm is an area used to determine a threshold for binarizing a multi-tone fingerprint image. This process is performed to reliably determine the threshold value in the area where the finger is placed.

FIG. 7 shows several specific examples of setting a binarized threshold value determination monitor dm using three finger detection monitors m0 to m4 in which a finger is detected. ing. For example, as shown in (a), when it is determined that a finger is placed on the finger detection monitors m0, m2, and m3, the binarization threshold determination monitor dm is set to the left side of the center, (B) As shown in FIG.
When it is determined that the finger is placed on 2, m3, and m4, the finger is set on the lower side from the center, and the finger is placed on the finger detection monitors m1, m2, and m4 as shown in (c). If it is determined that the finger is placed on the finger detection monitors m0, m1, and m2 as shown in FIG. It is set above the center. In this way, the binarization threshold determination monitor is set so that the finger is biased in a certain direction from the center.

In this setting, an area for determining a threshold is determined in advance in accordance with the state of each flag of the finger detection monitor, and information representing the area is stored in a memory in advance, and the flag of the finger detection monitor is obtained. When this is done, information indicating an area for determining the threshold value is read from the memory according to the state of the flag.

The same processing as described above is performed when the number of finger detection monitors determined to have a finger is four or five. Next, in step 202, step 201
A process for creating a histogram f (i) of pixels for each gray level of the image in the binarization threshold value determination monitor dm set in step (1) is performed. In the present embodiment, as shown in FIG.
Is divided into values from 0 to 255, and the closer to 0, the darker the image. The closer to 255, the brighter the image.

Next, at step 203, an initial value 0 is set to a variable j. Next, in step 204, the total number of pixels from i = 0 to i = j in the histogram shown in FIG. If the sum of the number of pixels exceeds the threshold K, in step 206, the value of j at that time is set as a binarization threshold, and a binary image is created using this threshold. If the total number of pixels does not exceed the threshold value K, j is incremented by 1 in step 205,
The process returns to step 203 to newly calculate the total number of pixels.

Here, when it is assumed that the thickness of the ridge line and the valley line of the fingerprint is equal, the threshold value K is half the total number of pixels of the binarization threshold value determination monitor. The threshold value K may be a numerical value set individually by calculating the thickness of the ridge line and the valley line of the registrant at the time of registration. Steps 105 and 10
6 functions as finger detecting means, step 110 functions as area determining means, step 112 functions as shape determining means, and step 206 functions as binarizing means.

As described above, according to the present embodiment, it is determined whether or not the area of the fingerprint image is sufficient to perform fingerprint collation using the finger detection monitor, and the area of the fingerprint image is determined. Even if it is sufficient to perform the collation, if the shape of the image is not appropriate, the input of the read fingerprint image is interrupted and the reading is performed again, so that the accuracy of the fingerprint collation can be improved. .

By setting the finger detection monitor, it is not necessary to calculate the entire fingerprint input screen (512 pixels × 512 pixels), so that the processing can be sped up. Further, in the process of converting a multi-tone image into a binary image, the binary process only needs to be performed once, and the process can be performed at high speed.

As shown in FIG. 7, the binarization threshold determination monitor is moved according to the state of the finger detection monitor determined to have a finger placed thereon, and the fingerprint is stored in the binarization threshold determination monitor. Since the boundary area is not included, the threshold can be reliably determined in the area where the finger is placed.

Although one embodiment of the present invention has been described above, the present invention is not limited to this embodiment and can be implemented in various modes. For example, the process of converting the multi-tone fingerprint image shown in FIG. 6 into a binary image may be performed as shown in the flowchart of FIG. Next, this will be described.

Step 301 is the same as step 201 described above, and a description thereof will not be repeated. Next, step 30
In 2, an initial value of the binarization threshold is set. This initial value may be a binarization threshold at the time of inputting a fingerprint image of the registrant in the past or a binarization threshold at the time of previous input.

Next, in step 303, the set binarization threshold value is compared with the value of each pixel, and a binarized image is created. For example, a pixel larger than the binarization threshold is set to black, and the other pixels are set to white. Next, in step 304, the total number of white or black pixels in the designated area is calculated. Next, in step 305, step 3
03 in order to determine whether or not the binarized image obtained in step 03 is the optimum binarized image, the sum is from Tmin to Tm.
ax is determined. Total value is Tmin
If it is between Tmax and Tmax, the binarized image is determined to be the optimal binarized image. If the binarized image is the optimum binarized image, the process ends. If it is not the optimum binarized image, the binarization threshold is changed in step 306, and the above processing is repeated.

[0034]

As described above, according to the present invention, it is determined whether or not the area of a fingerprint image is sufficient for performing fingerprint collation, and the shape of the fingerprint image is suitable for performing fingerprint collation. Is determined, and the binarized image is created only when the area of the fingerprint image is sufficient to perform fingerprint matching and the shape of the fingerprint image is appropriate for performing fingerprint matching. In addition, the accuracy of fingerprint collation can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a basic configuration of a fingerprint image input device according to the present invention.

FIG. 2 is a schematic block diagram of a fingerprint image input device according to one embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a memory map of the image memory shown in FIG. 2;

FIG. 4 is a flowchart illustrating a process of determining whether an input fingerprint image has a sufficient area and whether the shape of the fingerprint image is suitable for performing fingerprint collation based on an output of a finger detection monitor.

FIG. 5 is an explanatory diagram showing a case where the shape of a fingerprint image detected by a finger detection monitor is appropriate for performing fingerprint collation.

FIG. 6 is a flowchart illustrating a process of converting a multi-tone fingerprint image determined to be appropriate for fingerprint comparison into a binary image.

FIG. 7 is an explanatory diagram showing some specific examples of setting a binarization threshold determination monitor using a finger detection monitor.

FIG. 8 is an explanatory diagram showing a histogram of pixels for each gray level of an image in a binarization threshold determination monitor.

FIG. 9 is a flowchart illustrating another example of a process of converting a multi-tone fingerprint image into a binary image.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Fingerprint input part 2 ... Processing part 3 ... Fingertip 11 ...
Illumination 12 Prism 13 Image sensor 21 A / D converter 22
... Image memory 23 ... Calculation / control unit m0-m4 ... Finger detection monitor

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-63-244284 (JP, A) JP-A-63-61072 (JP, U) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06T 7/00 G06T 1/00

Claims (1)

(57) [Claims] 1. A fingerprint image input device for inputting a fingerprint image for performing fingerprint collation, comprising: a reading unit that optically reads a fingerprint from a subject's finger;
A fingerprint reading unit that obtains a fingerprint image signal; and, based on fingerprint image signals of a plurality of predetermined image regions of the fingerprint image obtained by the fingerprint reading unit, a finger is placed on a portion corresponding to each image region of the reading unit. Finger detection means for detecting whether or not the contact is made, and area determination means for determining whether or not the area of the fingerprint image is sufficient to perform fingerprint collation based on the output of the finger detection means. A shape discriminating unit for discriminating whether or not the shape of the fingerprint image is suitable for performing fingerprint collation based on an output of the finger detecting unit; and the area discriminating unit collating the fingerprint with the area of the fingerprint image. The fingerprint obtained by the fingerprint reading means in response to the determination that the fingerprint image is sufficient for Image signal A fingerprint image input device, comprising: a binarizing unit that converts the data into binary data using a fixed threshold value.