JP2009064261A - Finger movement position estimation device and fingerprint image processor using it, and finger movement estimation method and fingerprint image processing method using it - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a finger movement position estimation device and a fingerprint image processor using it, and a finger movement position estimation method and a fingerprint image processing method using it, for properly detecting a movement amount of a finger. <P>SOLUTION: This finger movement position estimation device 100 having a two line sensor 10 extending sideways, adjacently disposed lengthwise, and estimating a movement position and/or the movement amount of one pixel or below of the finger 150 sweeping on the two line sensor has: an estimation pixel data generation means 20 for generating estimation pixel data corresponding to the movement position and/or the movement amount of the finger of one pixel Bn present in a sweep direction among pixels of two rows and two columns based on pixel data obtained at first time t0 by the prescribed adjacent pixels of two rows and two columns; and a finger movement position estimation means 30 for comparing actual pixel data collected in the one pixel at second time t1 and the estimation pixel data generated by the estimation pixel data generation means, and estimating the movement position and/or the movement amount of the finger. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a finger movement position estimation apparatus, a fingerprint image processing apparatus using the same, a finger movement estimation method, and a fingerprint image processing method using the same, and more particularly to a finger that estimates a finger movement position using a two-line sensor. The present invention relates to a movement position estimation apparatus, a fingerprint image processing apparatus using the same, a finger movement estimation method, and a fingerprint image processing method using the same.

  2. Description of the Related Art Conventionally, when reconstructing an entire fingerprint image from partial fingerprint images, a method for estimating the amount of vertical and horizontal movement from the correlation by comparison between partial images is known (for example, see Patent Document 1). ).

A fingerprint image for obtaining a two-dimensional fingerprint image from a fingerprint image obtained by sweeping a finger on a one-dimensional sensor composed of two lines of a first line sensor and a second line sensor arranged at a distance. The fingerprint image is input to the first buffer and the second buffer, each of these buffers is divided, and a finger block is detected by detecting an image block determined to be the same from the correlation of the divided images. A fingerprint image input method is known in which a speed is obtained, distortion of the reduction or enlargement of the fingerprint image is detected from the speed, and the distortion is corrected by a predetermined method (see Patent Document 2).
JP-T-2005-53935 JP 2002-216116 A

  However, in the configuration described in Patent Document 1 described above, the pixel data serving as the comparison unit is actual data actually obtained from the sensor, and the comparison is not performed when compared with the comparison source data in units of one pixel. Therefore, movement of less than one pixel cannot be detected properly.

  In particular, in the case of a two-line sensor, since there is only two lines of actual data as a comparison source, the comparison in units of one pixel as shown in Patent Document 1 does not move in the horizontal direction. In other words, it is possible to determine only the state of moving one pixel or not, and it is not possible to detect the movement amount with high accuracy.

  Further, in the configuration described in Patent Document 2 described above, no consideration is given to a distortion including a lateral component, that is, a distortion when a lateral shift or a finger is tilted and swept.

  On the other hand, in the case of a sweep type sensor using a line sensor, on the other hand, lateral movement, that is, lateral deviation is relatively small with respect to vertical movement. For example, in the case of a two-line sensor, when the movement of one vertical line is completed, Usually, the movement is less than one pixel, and properly detecting this greatly affects the accuracy of image assembly and, consequently, the accuracy of fingerprint image authentication.

  Therefore, the present invention provides a finger movement position estimation device that appropriately detects the movement amount of a finger within one pixel, a fingerprint image processing device using the same, a finger movement position estimation method, and a fingerprint image processing method using the same. The purpose is to do.

In order to achieve the above object, a finger movement position estimating device (100) according to a first invention has a two-line sensor (10) extending in the horizontal direction and adjacently arranged in the vertical direction. A finger movement position estimation device (100) for estimating a movement position and / or movement amount of one finger or less of a finger (150) sweeping on a sensor (10),
Pixel data (An-1 (t0), An (t0), An (t0), the pixel data (An-1 (t0), An (t0), Bn-1 (t0), Bn (t0)), one pixel (Bn-1) in the sweep direction among the pixels in two rows and two columns (An-1, An, Bn-1, Bn) , Bn) estimated pixel data generating means (20) for generating estimated pixel data (Bn-1 (t1), Bn (t1)) corresponding to the movement position and / or movement amount of the finger (150).
Actual pixel data collected at the one pixel (Bn−1, Bn) at the second time (t1) and the estimated pixel data (Bn−1 (t1), Bn) generated by the estimated pixel data generating unit (T1)) and finger movement position estimation means (30) for estimating the movement position and / or movement amount of the finger.

  Thereby, the movement position of the finger of 1 pixel or less can be estimated appropriately, and it can be used for subsequent image processing.

According to a second aspect of the present invention, in the finger movement position estimating apparatus (100) according to the first aspect,
The estimated pixel data generation means (20) is configured to estimate pixel data (Bn-1 (t1), Bn-1 (t1), Bn (t1))
The second time (t1) is a time when the finger (150) has moved one line in the vertical direction on the two-line sensor (10),
The finger movement position estimation means (30) estimates the movement position and / or movement amount of the finger (150) in the lateral direction.

  Thereby, the movement position of the finger in the horizontal direction when the line data is updated can be estimated, and the movement position of the finger can be properly grasped.

According to a third aspect of the present invention, in the finger movement position estimating apparatus (100) according to the second aspect of the present invention,
The estimated pixel data generation means (20) is the estimated pixel data at a predetermined lateral movement position obtained by dividing the movement position of one pixel in the horizontal direction and the position moved by the one line in the vertical direction. (Bn-1 (t1), Bn (t1))
The finger movement position estimation means (30) compares the actual pixel data with estimated pixel data (Bn-1 (t1), Bn (t1)) at the predetermined lateral movement position, and performs the predetermined lateral movement. Among the estimated pixel data (Bn-1 (t1), Bn (t1)) at the position, the predetermined lateral movement having the pixel data (Bn-1 (t1), Bn (t1)) closest to the actual pixel data The position is estimated as a lateral movement position of the finger (150).

  Thereby, a plurality of lateral movement positions of a desired pixel or less as a reference can be selected and set, and it is possible to know which of these lateral movement positions is closest to the lateral movement position of the finger.

A fingerprint image processing apparatus (200) according to a fourth invention is the finger movement estimation apparatus (100) according to the third invention.
Each time the finger (150) moves on the two-line sensor (10) by one line in the vertical direction, the synthesis memory (90) records the frame data for the one line;
The horizontal movement position information representing the horizontal movement position is recorded in the synthesis memory (90) together with the frame data.

  As a result, the information on the lateral movement position of the finger can be recorded for each frame data, which can be used for subsequent image processing for image correction.

According to a fifth aspect of the present invention, in the finger movement position estimation apparatus (100) according to the third or fourth aspect of the present invention,
Further comprising a counting means for counting the frequency of the lateral movement position information representing the lateral movement position;
At the time when the sweep of the finger (150) is completed, a sweep inclination amount calculating means (40) for calculating the sweep inclination amount of the finger based on the frequency distribution of the lateral movement position information counted by the counting means. It is characterized by having.

  This is likely to occur when the finger is swept straight in the axial direction of the finger in an oblique state with respect to the line sensor, and the finger is swept with respect to the line sensor and the frame data image is inclined obliquely. When a fingerprint image has been formed, the amount of inclination can be calculated and used for subsequent image processing.

A fingerprint image processing apparatus (200) according to a sixth invention has a finger movement position estimating apparatus (100) according to the fourth invention,
It has lateral deviation correction means (50) for performing lateral deviation correction of the frame data recorded in the synthesis memory (90) using the lateral movement position information obtained by the finger movement position estimation device (100). To do.

  As a result, it is possible to correct the frame data image stored in the composite memory for the lateral displacement of the fingerprint image that is likely to occur when the finger is swept in the oblique direction with the finger vertically erected on the two-line sensor. Authentication accuracy can be improved.

A fingerprint image processing apparatus (200) according to a seventh invention includes the finger movement position estimation apparatus (100) according to the fifth invention,
The finger movement position estimation device (100) calculates the amount of sweep inclination of the finger (150),
A predetermined inclination correcting means (60) for correcting the inclination of the frame data recorded in the synthesis memory (90) using the sweep inclination amount of the finger is provided.

  Thereby, the inclination of the fingerprint image can be corrected, and the fingerprint authentication accuracy can be improved.

A finger movement position estimation method according to an eighth aspect of the present invention is the movement position of one pixel or less of the finger (150) that sweeps over the two line sensors (10) that extend in the horizontal direction and are adjacently arranged in the vertical direction. Or a finger movement position estimation method for estimating a movement amount,
Pixel data (An-1 (t0), An (t0), An (t0), the pixel data (An-1 (t0), An (t0), Bn-1 (t0), Bn (t0)), one pixel (Bn-1) in the sweep direction among the pixels in two rows and two columns (An-1, An, Bn-1, Bn) Bn) an estimated pixel data generating step for generating estimated pixel data corresponding to the movement position and / or movement amount of the finger (150);
The actual pixel data collected at the one pixel at the second time (t1) is compared with the estimated pixel data (Bn-1 (t1), Bn (t1)) generated by the pixel data generating means. And a finger movement position estimation step for estimating the movement position and / or movement amount of the finger (150).

  As a result, it is possible to estimate the movement position of a finger of one pixel or less, and it is possible to perform highly accurate image processing and the like using this.

A ninth invention is the finger movement position estimation method according to the eighth invention,
In the estimated pixel data generation step, estimated pixel data (Bn-1 (t1), Bn (t1)) corresponding to a position where the finger has moved one line in the vertical direction on the two-line sensor (10) is obtained. Generate
The second time is a time when the finger (150) has moved one line in the vertical direction on the two-line sensor (10),
In the finger movement position estimation step, the movement position and / or movement amount of the finger (150) in the lateral direction is estimated.

  This makes it possible to estimate the lateral movement position of the finger each time frame data is updated, which can be used for subsequent image correction.

A tenth aspect of the invention is the finger movement position estimation method according to the ninth aspect of the invention,
In the estimated pixel data generation step, estimated pixel data (Bn−) at a predetermined lateral movement position obtained by dividing the movement position of one pixel in the horizontal direction and the position moved by the one line in the vertical direction. 1 (t1), Bn (t1))
The finger movement position estimation step compares the actual pixel data with estimated pixel data (Bn-1 (t1), Bn (t1)) at the predetermined lateral movement position, and estimates at the predetermined lateral movement position. Among the pixel data (Bn-1 (t1), Bn (t1)), the predetermined lateral movement position having the pixel data (Bn-1 (t1), Bn (t1)) closest to the actual pixel data, It is estimated as a lateral movement position of the finger (150).

  Thereby, a plurality of lateral movement positions of a desired pixel or less as a reference can be selected and set, and it is possible to know which of these lateral movement positions is closest to the lateral movement position of the finger.

An eleventh aspect of the invention is the finger movement position estimation method according to the tenth aspect of the invention,
Each time the finger (150) moves on the two-line sensor (10) by one line in the vertical direction, the synthesis memory (90) records the frame data for the one line;
The horizontal movement position information representing the horizontal movement position is recorded in the synthesis memory (90) together with the frame data.

  This makes it possible to record the frame data and the information on the lateral movement position of the finger as a set, and when performing image processing such as image correction later, it can be performed based on appropriate lateral movement position information. it can.

A twelfth invention is the finger movement position estimation method according to the tenth or eleventh invention,
A counting step of counting the frequency of the lateral movement position information representing the lateral movement position;
A sweep inclination amount calculating step of calculating a sweep inclination amount of the finger based on a frequency distribution of the lateral movement position information counted in the counting step when the finger sweep is completed. .

  As a result, it is easy to occur when the finger is swept straight with respect to the line sensor and parallel to the axis of the finger. The fingerprint is swept with the finger tilted with respect to the line sensor and the frame data image is tilted diagonally. When an image has been formed, the amount of inclination can be calculated and used for subsequent image processing.

A fingerprint image processing method according to a thirteenth invention executes the finger movement position estimation method according to the eleventh invention,
A lateral shift correction step of performing lateral shift correction of the frame data recorded in the synthesis memory (90) using the lateral movement position information obtained by the finger movement position estimation method;

  As a result, it is possible to correct the frame data image stored in the composite memory for the lateral displacement of the fingerprint image that is likely to occur when the finger is swept in the oblique direction with the finger vertically erected on the two-line sensor. Authentication accuracy can be improved.

A fingerprint image processing method according to a fourteenth invention executes the finger movement position estimation method according to the twelfth invention,
By calculating the finger movement position estimation method, the amount of sweep inclination of the finger (150) is calculated,
A predetermined inclination correction step for correcting the inclination of the frame data recorded in the synthesis memory (90) using the sweep inclination amount of the finger.

  As a result, it is possible to perform image processing for appropriately correcting the inclination of the fingerprint image, correct the fingerprint image upright, and improve fingerprint authentication accuracy.

  Note that the reference numerals in the parentheses are given for easy understanding, are merely examples, and are not limited to the illustrated modes.

  According to the present invention, it is possible to appropriately estimate and grasp the movement position and / or movement amount of a finger within one pixel. Further, based on the estimated finger movement position and / or movement amount, a fingerprint image obtained by correcting the finger movement position and / or movement amount can be obtained, and the fingerprint image authentication accuracy can be improved.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

  FIG. 1 is a functional block diagram showing a schematic configuration of a finger movement position estimation apparatus 100 and a fingerprint image processing apparatus 200 according to an embodiment to which the present invention is applied.

  In FIG. 1, the finger movement position estimation apparatus 100 according to the present embodiment includes a two-line sensor 10, an estimated pixel data generation unit 20, and a finger movement position estimation unit 30. Moreover, the finger movement position estimation apparatus 100 may be provided with the synthetic | combination memory 90, the sweep inclination amount calculation means 40, and the count means 45 as needed.

  The fingerprint image processing apparatus 200 may further include a lateral deviation correction unit 50 and an inclination correction unit 60 in addition to the finger movement position estimation apparatus 100.

  Then, the finger movement position estimation device 100 and the fingerprint image processing device 200 may include a support base 80 that supports the two-line sensor 10 and places the finger 150 to sweep.

  The two-line sensor 10 is a fingerprint image collecting means having two line sensors 11 and 12 of an A-line line sensor 11 and a B-line line sensor 12. The line sensors 11 and 12 are sensors for capturing and collecting a line-like fingerprint image. Each of the line sensors 11 and 12 includes only one pixel in the vertical direction, but includes a plurality of pixel columns in the horizontal direction. For example, the line sensors 11 and 12 have a unit level of 100 units such as 126 pixels, 252 pixels, and 512 pixels. Pixels are arranged.

  The two-line sensor 10 has a configuration in which two line sensors 11 and 12 extend in the horizontal direction and are adjacently arranged in the vertical direction. The finger 150 sweeps and moves on the two-line sensor 10 in the vertical direction. At this time, the two line sensors 11 and 12 each collect a linear partial image of the fingerprint of the finger 150. Then, the partial images are combined so as to be stacked in the vertical direction to form an entire fingerprint image. This is used, for example, for fingerprint authentication.

  In this embodiment, the two-line sensor 10 having only two line sensors 11 and 12 will be described as an example. However, if the sensor has two or more line sensors, two desired sensors are used. This embodiment can be applied by selecting the line sensor. Therefore, the present embodiment can be suitably applied to a plurality of line sensors including two line sensors 10 such as three lines and four lines. Even when two line sensors of such a plurality of line sensors are used, Examples are included.

  The finger 150 is generally swept from the A-line line sensor 11 to the B-line line sensor 12 in FIG. 1, that is, the finger 150 is swept forward, but in the opposite direction. Also good. Hereinafter, in the present embodiment, the case where the finger 150 is swept to the front side will be described as an example. However, if the relationship between the A line and the B line is reversed, the present embodiment is independent of the sweep direction of the finger 150. Can be applied.

  In the finger movement estimation apparatus 100 according to the present embodiment, the line sensors 11 and 12 include predetermined two rows and two columns of adjacent pixels An-1, An, Bn-1, and Bn for finger movement determination. These pixels may be those previously selected for finger movement position estimation, and have at least four pixels of 2 rows and 2 columns adjacent to each other. In the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment, pixel data of the finger movement position estimation pixels An-1, An, Bn-1, and Bn obtained at an arbitrary time are used. Among these pixels, pixel data of one pixel in the sweep direction of the finger 150 is estimated and expressed. For example, if the finger 150 sweeps to the right front side, the pixel data of the pixel Bn in the sweep direction is represented by the pixel data of the pixels An-1, An, Bn-1, and Bn for finger movement position estimation. When sweeping to the left front side, pixel data of the pixel Bn-1 in the sweep direction is expressed.

  The number of pixels An-1, An, Bn-1, and Bn for estimating the finger movement position may be any as long as it is four pixels or more, and a set of four pixels may be provided at a plurality of positions in the horizontal direction. The finger movement position may be estimated using all the pixels of the two-line sensor 10.

  The estimated pixel data generation unit 20 is an arithmetic processing unit that estimates and generates pixel data of one pixel in the sweep direction of the finger 150. Although details of the specific processing contents will be described later, one pixel in the sweep direction is obtained by using pixel data of finger movement position estimation pixels An-1, An, Bn-1, and Bn according to a predetermined arithmetic expression. Represents pixel data.

  Note that the estimated pixel data generation means 20 may include a memory 21 therein. Using the pixel data of the pixels An-1, An, Bn-1, and Bn for finger movement position estimation acquired at an arbitrary time (hereinafter referred to as “first time”) in the memory 21, another time is used. In order to express the pixel data of one pixel at (hereinafter referred to as “second time”), the pixel data of the pixels An−1, An, Bn−1, and Bn acquired at the first time are stored in the memory 21. You may make it memorize.

  The finger movement position estimation means 30 is an arithmetic processing means for estimating the movement position of the finger 150, and is sampled at the second time with estimated pixel data of one pixel in the sweep direction generated by the estimated pixel data generation means 20. It is a means for comparing the actual pixel data and estimating the movement position of the finger 150. For example, when there are a plurality of estimated pixel data of one pixel in the sweep direction corresponding to a plurality of estimated movement positions, the finger movement position estimation unit 30 uses the actual pixel data and the plurality of estimated pixel data as a pattern. If matching is performed and the estimated pixel data closest to the actual pixel data can be identified, the movement position of the finger 150 can be estimated to be at a position corresponding to the identified estimated pixel data. For example, the finger movement position estimation unit 30 may estimate the finger movement position using such pattern matching.

  The composite memory 90 is a recording means for writing and recording line-shaped partial image data collected by the line sensors 11 and 12 of the two-line sensor 10 as frame data every time one line movement is completed. Therefore, the frame data is accumulated in the synthesis memory 90, and finally the entire image of the synthesized fingerprint is formed.

  The synthesis memory 90 can record information related to the frame data in addition to the linear image data of the frame data. Although details will be described later, in the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment, when the frame data is written in the synthesis memory 90, finger movement position information of less than one pixel is also recorded. As a result, the finger movement position information is recorded every time the frame data is written. By accumulating the finger movement position information, the final finger movement position when the sweep of the finger 150 is completed can be estimated. Then, based on the finger movement position information in the final or mid-sweep stage, necessary correction processing can be performed when a composite image of the entire fingerprint is formed. Thereby, a fingerprint image can be brought close to an actual fingerprint, and fingerprint authentication accuracy can be improved.

  When the finger movement position estimation in the finger movement position estimation unit 30 is performed by pattern matching with a plurality of estimated pixel data, the counting unit 45 determines which finger movement position and how many times with respect to the estimated finger movement position information. This is a means of counting the estimated frequency. That is, since the finger 150 moves by sweeping, the number of data of the finger movement position information increases and accumulates with the sweep. The counting means 45 counts how many times each finger movement position is estimated in the entire finger movement position information data, and records the frequency distribution. The counting means may be, for example, a counter set for each finger movement position, may count the movement position when frame data is recorded by the finger movement position estimation means 30, or may be combined memory 90. It is also possible to read the finger movement position information recorded in, and to count the frequency from this.

  The sweep inclination amount calculation means 40 is an arithmetic processing means for calculating the inclination amount of the sweep from the finger movement position information stored and recorded in the synthesis memory 90. Since the movement direction and movement amount of the finger 150 can be calculated from the accumulated movement position information of the finger 150, the inclination amount of the entire sweep can be calculated. The sweep tilt amount calculation means 40 performs this calculation and appropriately estimates the sweep tilt amount of the finger 150.

  The lateral deviation correction means 50 is an image processing means for monitoring lateral movement position information of the finger movement position information recorded in the composite memory 90 and performing lateral deviation correction processing on the frame data when there is a lateral deviation. As described above, when the frame data is recorded, the finger movement position information is also recorded in the synthesis memory 90. However, the frame data includes the line sensor 11 when the finger 150 moves one line in the vertical direction, Since it means 12 line image data, it is the finger movement position information in the horizontal direction that is actually recorded as the finger movement position information. Therefore, the lateral shift of the frame data can be detected by monitoring the lateral finger movement position information recorded in the composite memory 90. Therefore, when the lateral shift of the frame data is detected, the lateral shift correction unit detects this. Image processing to correct

  These corrections may be performed after the recording of all the frame data is completed, or may be performed as needed, for example, by correcting one pixel movement when the lateral shift is accumulated and reaches one pixel. .

  The inclination correction means 60 is a means for performing image processing for correcting the sweep inclination amount when the sweep inclination amount calculation means 40 calculates the sweep inclination amount. For example, a correction process may be performed so that the fingerprint image is erected by performing rotation correction that corrects the amount of tilt on the entire fingerprint image having the tilt. The tilt correction means 60 may be applied with various image processing that corrects the tilt amount of the image and erects the image.

  The estimated pixel data generation means 20, the finger movement position estimation means 30, the sweep inclination amount calculation means 40, the lateral deviation correction means 50, and the inclination amount correction means 60 may all be configured as an arithmetic processing means 70 that performs arithmetic processing. For example, it may be realized by a computer that operates according to a predetermined program, a predetermined electronic circuit, an ASIC, or the like.

  Among the calculation processing means 70, the estimated pixel data generation means 20, the finger estimated position estimation means 30 and the sweep inclination amount calculation means 40 are calculation processing means for estimating the movement position of the finger 150, and include count means 45. In addition, the finger movement position estimation calculation processing means 71 is configured. Further, the lateral deviation correction unit 50 and the inclination amount correction unit 60 are image processing units that perform image correction, and constitute an image correction processing unit 72. The finger movement position estimation calculation processing means 71 and the image correction processing means 72 may be configured as an integral calculation processing means, or may be configured as independent separate means. These can take various modes depending on the application.

  Next, the contents of the estimated pixel data generation calculation executed by the estimated pixel data generation unit 20 will be described with reference to FIG. FIG. 2 is a diagram for explaining the estimated pixel data generation calculation performed by the estimated pixel data generation unit 20.

  FIG. 2A shows pixel data of pixels An−1, An, Bn−1, and Bn for finger movement position estimation in predetermined two rows and two columns adjacent to the two line sensor 10 at the first time t = t0. FIG. In FIG. 2A, four pixels An−1, An, Bn−1, and Bn in 2 rows and 2 columns of the line sensors 11 and 12 that are located in the substantially central portion of the finger 150 have a first time t = t0. The pixel data is respectively acquired in.

  FIG. 2 (b) shows a predetermined 2-row 2-column finger movement position estimation pixel An-1 adjacent to the 2-line sensor 10 in a state in which the finger 150 is moved by the sweep at the second time t = t1. , An, Bn−1, and Bn pixel data. In FIG. 2B, the finger 150 is swept to the right front side (lower right side), and the pixels An-1, An, Bn-1, and Bn for estimating the finger movement position in 2 rows and 2 columns are relatively fingertips. 150 shows a state of moving slightly to the upper left of 150.

  That is, the finger 150 moves Δy to the near side (downward) in the vertical direction, and Δx moves to the right side in the horizontal direction. The two-line sensor at the second time t = t1 is a thick solid line. The position under 10 fingers 150 is shown. On the other hand, the position of the two-line sensor 10 under the finger 150 at the first time t = t0 is indicated by a thin solid line. The finger region located on the two-line sensor 10 at the first time t = t0 moves Δx to the right and Δy to the lower at the second time t = t1, thereby causing the finger 150 to move. It can be seen that the relative position is moved. However, Δx and Δy are movement amounts in terms of one pixel unit, and both Δx = 1 and Δy = 1 mean movement by one pixel.

  Here, the pixel data of the pixel Bn observed at the second time t = t1 is a combination of the pixel data of each pixel An-1, An, Bn-1, and Bn observed at the first time t = t0. It turns out that it is. Then, the combination ratio of the pixel values on the line sensors 11 and 12 observed at the first time t = t0 in one pixel Bn in the sweep direction on the line sensor 12 at the second time t = t1. Is expressed as follows using the movement amounts Δx, Δy of the finger 150 from the area ratio.

よって、これらから、第２の時刻ｔ＝ｔ１時の画素Ｂｎの画素値Ｂｎ（ｔ１）は、以下のように表せる。

Therefore, from these, the pixel value Bn (t1) of the pixel Bn at the second time t = t1 can be expressed as follows.

式（１）に示したように、第１の時刻ｔ＝ｔ０において取得した指移動位置推定用の画素Ａｎ−１、Ａｎ、Ｂｎ−１、Ｂｎの画素データＡｎ−１(ｔ０)、Ａｎ(ｔ０)、Ｂｎ−１(ｔ０)、Ｂｎ(ｔ０)を用いて、第２の時刻ｔ＝ｔ１におけるスイープ方向にある画素Ｂｎの画素値を表現することができる。推定画素データ生成手段２０においては、（１）式に従った演算を行ない、スイープ方向にある１画素を、第１の時刻ｔ＝ｔ０における自己を含む周辺の画素の画素データを用いて表現する。その際、第１の時刻ｔ＝ｔ０で取得した周辺の指移動位置推定用の画素Ａｎ−１、Ａｎ、Ｂｎ−１、Ｂｎの画素データＡｎ−１(ｔ０)、Ａｎ(ｔ０)、Ｂｎ−１(ｔ０)、Ｂｎ(ｔ０)は、メモリ２１に記憶しておくようにしてもよい。

As shown in the equation (1), the pixel data An−1, An, Bn−1, Bn pixel data An−1 (t0), An () for the finger movement position estimation acquired at the first time t = t0. The pixel value of the pixel Bn in the sweep direction at the second time t = t1 can be expressed using t0), Bn-1 (t0), and Bn (t0). The estimated pixel data generation means 20 performs an operation according to the equation (1), and expresses one pixel in the sweep direction using pixel data of surrounding pixels including itself at the first time t = t0. . At that time, the pixel data An-1 (An), An (t0), An (t0), Bn− of the surrounding finger movement position estimation pixels An−1, An, Bn−1, Bn acquired at the first time t = t0. 1 (t0) and Bn (t0) may be stored in the memory 21.

  According to the equation (1), the pixel data Bn (t1) of the pixel Bn at the second time t = t1 is the surrounding finger movement position estimation pixels An-1, An, Bn− at the first time t = t0. 1 and Bn pixel data An-1 (t0), An (t0), Bn-1 (t0), Bn (t0) can be expressed by specifying Δx and Δy in equation (1), A pixel value of the pixel Bn (t1) at a desired position can be generated.

  In FIG. 2, the example in which the finger 150 is moved to the right front (lower right) has been described. However, when the finger 150 is moved to the left front (lower left), the pixel data Bn−1 of the pixel Bn−1 is described. (T1) is generated. When the finger 150 is moved to the upper right, the pixel data An (t1) of the pixel An, and when the finger 150 is moved to the upper left, the pixel data An− of the pixel An−1 is generated. 1 (t1) is generated.

  FIG. 3 is a diagram for explaining the calculation processing contents of the estimated pixel data generation means 20 when the finger movement position estimation pixels of the two-line sensor 10 are six pixels.

  In FIG. 3, there are three finger movement position estimation pixels on the line sensor 11 of the A line, An-1, An, and An + 1, and pixels on the line sensor 12 of the B line are Bn-1, Bn. , Bn + 1, which is different from FIG. 2 in that it is composed of six adjacent pixels.

  In this case, the estimated pixel data can be generated for the pixel Bn whether the finger 150 moves to the lower right or the lower left. That is, when the finger 150 moves in the lower right direction, the pixel data An−1, An, Bn−1, Bn of the pixel data An−1 (t0), An (t0), An at the first time t = t0, The pixel data Bn (t1) of the pixel Bn at the second time t = t1 is expressed using Bn−1 (t0) and Bn (t0). As a result, the pixel data Bn (t1) can be expressed using the equation (1).

  On the other hand, when the finger 150 moves to the lower left, the pixel data An (t0), An + 1 (t0) of the pixels An, An + 1, Bn, Bn + 1 at the first time t = t0, as shown in Expression (2). , Bn (t0), Bn + 1 (t0), the pixel data Bn (t1) of the pixel Bn at the second time t = t1 can be expressed.

このように、推定画素データを生成する対象を画素Ｂｎに固定し、指１５０のスイープ方向に応じて、画素Ｂｎの画素データを表現する周辺の画素の組合せを変更するようにしてもよい。推定画素データ生成手段２０で用いる指移動位置推定用画素は、用途に応じて、２ラインセンサ１０上の任意の画素を利用することができる。

As described above, the target for generating the estimated pixel data may be fixed to the pixel Bn, and the combination of the surrounding pixels expressing the pixel data of the pixel Bn may be changed according to the sweep direction of the finger 150. As the finger movement position estimation pixel used in the estimated pixel data generation unit 20, any pixel on the two-line sensor 10 can be used depending on the application.

  Conversely, when the finger 150 sweeps to the lower right, the estimated pixel data Bn + 1 (t1) of the pixel Bn + 1 at the second time t = t1 is converted into the surrounding pixel data An ( t0), An + 1 (t0), Bn (t0), Bn + 1 (t0), and when the finger 150 sweeps to the lower left, the estimated pixel data Bn-1 (t1) at the second time of the pixel Bn-1 ) Is expressed by the peripheral pixel data An-1 (t0), An (t0), Bn-1 (t0), Bn (t0) at the first time t = t0, and the adjacent 6 are adjacent to each other depending on the sweep direction of the finger 150. The estimated pixel data may be generated for the pixel data Bn−1 (t0) and Bn + 1 (t0) of the pixels Bn−1 and Bn + 1 located at the ends of the pixels.

  As described above, various combinations can be applied to the combination of the target pixel for generating the estimated pixel data and the peripheral pixel serving as the original data, and can be variously changed and set depending on the mode of the two-line sensor 10.

  Next, the contents of the arithmetic processing performed by the finger movement position estimation unit 30 will be described with reference to FIGS. 4 and 5.

  FIG. 4 is a diagram to be compared for explaining calculation processing performed by the finger movement position estimation unit 30 of the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment.

  FIG. 4A is a diagram showing pixel data A1, A2, A3, B1, B2, and B3 of six adjacent pixels on the two-line sensor 10 at the first time t = t0.

  FIG. 4B is a diagram illustrating pixel data of the two-line sensor 10 in a state where the finger 150 moves vertically one line at the second time t = t1 and there is no lateral movement. . In FIG. 4A, pixel data A1, A2, and A3 of the line sensor 11 of the A line are moved directly below the line sensor 12 of the B line. In the description so far, Δx = 0 and Δy = 1. In this state, if attention is paid to the pixel data of the line sensor 12 of the B line, it can be seen that the pattern coincides with the pixel data of the line sensor 11 of the A line in the state of FIG. It is possible to detect that one line has moved in the vertical direction and no movement in the horizontal direction. The pixel data of the A-line line sensor 11 is new pixel data.

  FIG. 4C shows pixel data of the two-line sensor 10 in a state where the finger 150 is moved by 45 ° in the lower right direction and one line in the vertical direction at the second time t = t1. It is. In terms of coordinate variables, Δy = 1 and Δx = 1. The pixel data A1, A2, A3 on the A-line line sensor 11 in FIG. 4A are shifted to the right by one pixel on the B-line line sensor 12 in FIG. You can see that it is moving in the direction. As described above, when the finger 150 sweeps at an inclination of 45 °, the pixel data acquired at the first time t = t0 by the line sensor 11 and the line sensor 12 sampled when moving one line in the vertical direction. By comparing the pixel data, it can be seen that the finger 150 is tilted to the right at 45 ° and moved laterally by one pixel.

  As described above, when only the actual pixel data actually captured and acquired is used in the two-line sensor 10, it is possible to detect the movement in the horizontal direction only with respect to the inclination of 0 ° and the inclination of 45 °. In addition, it is possible to determine whether the pixel data collected by the line sensor 11 is closer to 0 ° inclination pixel data or 45 ° inclination pixel data, and to estimate that the pixel data is inclined 0 ° or 45 °. .

  However, since only 0 ° and 45 ° can be determined, it does not make much sense in a fingerprint image input device that hardly tilts by 45 ° and sweeps.

  Therefore, in the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment, a plurality of estimated pixel data when moving a predetermined amount of one pixel or less in the horizontal direction using the estimated pixel data generation means 20 are used. Even if it is a movement of 1 pixel or less, an appropriate amount of lateral movement can be obtained by determining which of the estimated pixel data the pixel data actually prepared and actually collected is close to. It is possible to estimate.

  FIG. 5 is a diagram illustrating an example of pixel data of the two-line sensor 10 when the finger 150 moves in the horizontal direction by one pixel or less when the finger 150 moves in one line in the vertical direction.

  FIG. 5A is a diagram illustrating pixel data of the line sensor 12 when the finger 150 is moved one vertical line with the finger 150 tilted by 22.5 °. This is the case where the movement amount is Δx = 0.5 and Δy = 1, and the movement amount is half a pixel in the horizontal direction with respect to the movement of one pixel in the vertical direction. By substituting Δx = 0.5 and Δy = 1 into equation (1), the pixel data B2 can be expressed using the pixel data A1 and A2, and the pixel data A2 and A3 are used to represent the pixel data. Data B3 can be expressed.

  FIG. 5B is a diagram showing pixel data of the line sensor 12 when the finger 150 moves one vertical line while being tilted by 11.25 °. This is pixel data when the case of 22.5 ° in FIG. 5A is further divided in half, and pixel data when Δx = 0.25 and Δy = 1.

  FIG. 5C is a diagram illustrating pixel data of the line sensor 12 when the finger 150 moves one line in the vertical direction with the finger 150 tilted by 33.75 °. This is pixel data when the case of 22.5 ° in FIG. 5A is further divided in half and lateral movement is increased, and is pixel data when Δx = 0.75 and Δy = 1. .

  As described above, at the first time t = t0, the second time t = t1 when the finger 150 moves in the vertical direction by one line using the pixel data A1, A2, and A3 obtained by the line sensor 11. In this example, the estimated pixel data when the horizontal direction of the line sensor 12 moves by one pixel or less is generated, and these are compared with the actual pixel data actually collected, and the most similar estimated pixel data is selected and specified. An appropriate lateral movement position can be estimated. That is, if the actual pixel data is most similar to the pixel data having the inclination of 22.5 °, it can be estimated that the finger position is moved at the inclination of 22.5 °, and the pixel having the inclination of 11.25 ° is obtained. If it is the most similar to the data, it can be estimated that the finger 150 is moving at an inclination of 11.25 °.

  The finger movement position estimation means 30 performs a finger movement position estimation calculation using such pattern matching. The comparison between the plurality of estimated pixel data and the actual pixel data is performed by, for example, calculating a difference between pixel values of pixels to be compared, and determining the estimated pixel data having the smallest value as the most similar pixel data. Also good.

  The estimated pixel data of the line sensor 12 can be generated for an arbitrary amount of lateral movement obtained by dividing the lateral movement of one pixel. In FIGS. 5A to 5C, the example in which one pixel is equally divided into four has been described for ease of calculation. However, estimated pixel data may be generated for a different angle. The estimation accuracy may be increased by increasing the number of generations of estimated pixel data.

  FIG. 5D is a diagram showing pixel data of the line sensor 12 when the finger 150 moves one vertical line with the finger 150 tilted by 9 °. This is pixel data when 45 ° is divided into 1/5, and is pixel data when Δx = 0.2 and Δy = 1. As described above, the movement position for generating the estimated pixel data can be variously set according to the application.

  4 and 5, for the sake of easy explanation, an example in which the estimated pixel data is generated at the stage of moving one line in the vertical direction has been described. However, the amount of movement Δy in the vertical direction is Even in cases other than 1, it is possible to generate estimated pixel data. However, when the finger 150 moves one line in the vertical direction, it is common to collect the frame data and update the data. Therefore, the horizontal direction is based on the time point Δy = 1 when the line 150 is moved. A method of estimating the movement position is simple and preferable.

  It should be noted that the vertical movement of one line, that is, the frame data update time, is not only detected by using the equation (1), but the movement of one line on the basis of when Δy = 1, and by other methods. One line movement may be detected. Regarding the one-line movement detection, there is an advantage that the calculation processing burden of the finger movement position estimation means 30 can be reduced by combining other methods. Specific examples of other methods will be described later.

  The frame data is updated by the finger movement position estimating means 30 recording pixel data for one line acquired by the line sensors 11 and 12 in the synthesis memory 90. Thereby, a fingerprint image is formed. In the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment, when the frame data is updated, information on the lateral movement position is recorded together with the recording of the frame data. Image data can be corrected later using the accumulated data.

  FIG. 6 shows lateral movement position information that is written when the finger movement position estimation means 30 writes frame data to the composite memory 90. In FIG. 6, the lateral movement position information is shown in a table format.

  In the table of FIG. 6, Index, Δx, Δy, angle, and frequency are listed as items constituting the table. Index is an index indicating the amount of movement in the horizontal direction, that is, the amount of lateral deviation. In this embodiment, the amount of movement in the vertical direction is set to Δy = 1, and the amount of movement Δx in the horizontal direction is Since it is equally divided into four in both positive and negative directions, it has the same meaning as Δx. The angle means the amount of inclination of the sweep of the finger 150 described in FIG. 5, and this is expressed as a positive or negative angle. This is also proportional to the amount of movement Δx in the horizontal direction.

  In the table of FIG. 6, the frequency is, for example, the number of times of the lateral movement position determined every time the frame data is updated and determined as the maximum similarity by pattern matching. That is, the finger movement position estimation unit 30 performs pattern matching on the estimated pixel data having the index values of −4 to 4 with respect to the actual pixel data collected by the two-line sensor 10. Then, the index determined to be the most similar to this is recorded in the synthesis memory 90 together with the frame data as an index representing the amount of lateral deviation, but the number of times is recorded as a frequency, and finally counted. Counted by means 45. In addition, the sign of the sign means that the negative sign means that the image is laterally shifted to the right, and the positive sign means that the image is laterally shifted to the left.

  In FIG. 6, the frequency is one of the items of the lateral movement position information written simultaneously with the recording of the frame data, but the frequency does not necessarily need to be recorded or updated simultaneously with the recording of the frame data. That is, at the same time as the frame data, for example, only Index may be accumulated and recorded as the lateral movement position information, and the count may be counted by the counting means 45 when the sweep is completed. Since the frequency only needs to know the total number of estimated horizontal movement positions, it is not always necessary to update the frame data. Finally, if the horizontal movement position information data is read and the frequency is counted, This is because the frequency distribution can be calculated.

  On the contrary, when only the frequency is required, the counting means 45 is sequentially counted when the lateral movement position is estimated by the finger movement position estimation means 30 regardless of the writing of the frame data into the synthesis memory 90. The horizontal movement position selected in (4) may be counted up. As will be described in detail later, for example, when calculating the inclination angle of the sweep of the finger 150, the correspondence with the frame data is unnecessary, and only the final frequency distribution is required. The frequency of the counting means 45 may be updated directly from the position estimating means 30.

  The finger movement position estimation means 30 performs writing of the table relating to such a lateral movement position together with the update of the frame data, thereby performing image correction using the lateral movement position information during the entire image formation. It can be performed.

  Next, the content of the image processing for lateral misalignment image correction executed by the lateral misalignment correcting unit 50 will be described with reference to FIG.

  In FIG. 6, the index is accumulated and recorded every time frame data is updated. For example, when the accumulation result is -4 or less, the index is shifted to the left by one pixel, and when it is 4 or more. The lateral shift can be corrected by forming the entire image while shifting it horizontally by one pixel to the right. In other words, since the minimum amount of lateral displacement detection is an amount obtained by dividing one pixel into four, the time when the displacement amount cumulative value becomes 4 corresponds to one pixel displacement in pixel conversion. In the image correction, since movement correction is performed in units of pixels arranged in the synthesis memory 90, lateral shift correction is performed when the accumulated value exceeds a multiple of 4.

  In this way, the index indicating the lateral shift in the synthesis memory 90 written by the finger movement position estimating unit 30 is accumulated, and the lateral shift is corrected when the lateral shift amount reaches the pixel unit, thereby preventing the lateral shift. The authentication accuracy of the entire fingerprint image can be improved.

  Next, the contents of the arithmetic processing executed by the sweep inclination amount calculation means 40 will be described with reference to FIG. In FIG. 6, there is a frequency item indicating the frequency of the Index, which is counted by the counting means 45, and the sweep inclination amount calculating means 40 calculates the average value from the histogram based on this Index. Then, the inclination angle can be estimated from the inclination angle-histogram average value correspondence table obtained as an actual measurement value in advance. For example, in the example of FIG. 6, Index1 is frequency 12, Index2 is frequency 3, Index0 is frequency 2, and Index3 is frequency 1, and the average value of the histogram is about 1.16. From this value, the tilt angle can be calculated using the above-described tilt angle-histogram average value correspondence table.

  For ease of understanding, the frequency shown in the table of FIG. 6 has been described. However, as described above, the frequency may be sequentially counted from the finger movement position estimation unit 30 by the counting unit 45. Alternatively, the lateral movement position information without the frequency item recorded in the synthesis memory 90 may be read, and the frequency may be counted from this time. Various modes can be applied to these depending on applications.

  As described above, the sweep inclination amount calculating unit 40 can calculate the sweep inclination amount based on the frequency distribution of the lateral movement position information counted by the counting unit 45.

  Further, the inclination correction unit 60 performs the correction process for the entire image in the synthesis memory 90 based on the inclination angle calculated by the sweep inclination amount calculation unit 40. The inclination correction means 60 may be executed by various predetermined methods. For example, the inclination angle may be corrected by a rotation matrix that rotates the inclination angle or a map that corrects the inclination. By performing such correction, the fingerprint image authentication accuracy can be improved.

  FIG. 7 is a diagram illustrating an example of the sweep mode of the finger 150. In FIG. 7, the finger 150 has the inclination of the center axis of the finger itself kept perpendicular to the two-line sensor 10 (not shown), but the sweep has caused a lateral shift. The lateral shift correction means 50 of the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment can perform image correction particularly effectively for such lateral shift.

  FIG. 8 is a diagram illustrating an example of a sweep mode of the finger 150 different from that in FIG. 8, the sweep direction of the finger 150 is kept straight, but since the finger 150 is in contact with the two-line sensor 10 (not shown), the sweep is performed in the tilted state. ing. The tilt correction means 60 of the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment can perform image correction particularly effectively for such a lateral shift.

  As described above, the lateral deviation correction by the lateral deviation correction means 50 and the inclination correction by the inclination correction means 60 are different in appropriate correction means depending on the sweep mode. Therefore, the lateral deviation correction unit 50 and the inclination correction so as to be compatible with both sweep modes. Both of the means 60 may be provided, or only one having a particularly high frequency may be provided. These may be appropriately combined depending on the sweep mode.

  Next, a finger movement position estimation method and an image processing method executed by the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment will be described with reference to FIG. FIG. 9 is a flowchart illustrating a processing flow of the finger movement position estimation method and the image processing method according to the present embodiment.

  In FIG. 9, in step 100, frame data at the first time t = t0 is acquired by the two-line sensor 10 having the line sensors 11 and 12. Among the acquired frame data, the pixel data for finger movement position estimation is stored in the memory 21 of the estimation pixel data generation unit 20, and the estimation pixel data generation unit 20 performs a plurality of predetermined lateral shift positions (lateral movement positions). ) Is generated.

  In step 110, it is determined whether or not the finger 150 has moved one line. In Expression (1), Δy = 1 may be substituted, and determination may be made based on whether or not the pixel data collected by the line sensor 12 matches the pixel data acquired by the line sensor 11 at the first time. However, it may be determined by other methods described later. When one line has not been moved, the process returns to step 100 and the acquisition of frame data is repeated. If it has moved one line, the process proceeds to step 120.

  In step 120, the estimated pixel data estimated and generated at the plurality of lateral displacement positions generated in step 100 by the finger movement position estimation means 30 is compared with the actual pixel data collected by the two-line sensor 10, and the maximum similarity is obtained. Search for estimated pixel data indicating a point. The lateral movement position corresponding to the estimated pixel data indicating the maximum similarity is set as the lateral shift amount of the frame data at the second time t = t1. For example, in the example shown in FIG. 6, the lateral deviation amount is −4 to 4. The lateral shift amount is recorded in the synthesis memory 90.

  In step 130, the frequency of the lateral movement position information is updated. The frequency counted by the counting unit 45 may be updated from the finger movement position estimation unit 30, or the frequency data recorded in the synthesis memory 90 may be updated. Further, when the lateral movement position information is read from the composite memory 90 after the sweep is completed and the frequency is updated at that time, step 130 is executed between step 150 and step 160.

  In step 140, the line is updated and the frame data is recorded in the synthesis memory 90.

  In step 150, it is determined whether or not the sweep is finished. When the sweep is not completed, the process returns to step 100 and the acquisition of frame data is repeated. On the other hand, when the sweep is finished, the routine proceeds to step 160.

  In step 160, the inclination amount calculation means 40 calculates an average value from the frequency distribution counted by the counting means 45, and estimates the inclination angle during the sweep using the correspondence table.

  In step 170, image correction is executed by the lateral deviation correction means 50 and / or the inclination correction means 60, and the process ends. In step 170, image processing is performed to bring the fingerprint image closer to the actual fingerprint, thereby improving the accuracy of fingerprint authentication.

  Note that steps 160 and 170 are not essential in the finger movement position estimation method, and only necessary processing may be executed in the fingerprint image processing method.

  Next, another method for detecting the movement of one line in the vertical direction will be described with reference to FIGS. 10 and 11.

  FIG. 10 is a diagram illustrating a schematic configuration of a line movement determination unit 300 applied to the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment. In FIG. 10, main components of the line movement determination unit 300 include a two-line sensor 10 and an arithmetic control unit 160. The arithmetic control unit 160 includes a reference image storage unit 120, an image difference calculation unit 130, and a partial image data update determination position setting unit 140. The constituent elements of the arithmetic control means 160 may be constituted by various arithmetic means like a computer or the like. The two-line sensor 10 may be provided on the support base 80 similarly to the finger movement position estimation device 100. Similarly to the finger movement position estimation apparatus 100, the line movement determination unit 300 includes a synthesis memory 90 for synthesizing images and generating an image of the entire fingerprint. The two-line sensor 10 has two line sensors 11 and 12 as described above.

  Hereinafter, the line sensor 11 will be referred to as the first line sensor 11 and the line sensor 12 will be referred to as the second line sensor 12 in order to facilitate the distinction between the line sensors 11 and 12.

  The reference image storage unit 120 is a unit for storing a fingerprint line image acquired by the second line sensor 12 at a predetermined time. Any storage means such as a rewritable memory may be used. For example, a volatile memory or the like may be used. Although details will be described later, the reference image storage unit 120 stores the line image acquired by the second line sensor 12 at a predetermined time as a reference line image, thereby making this a reference image for comparison reference. By comparison with this, it is possible to make a determination such as movement of one line.

  The image difference calculation means 130 is an arithmetic means for comparing two line images and calculating the image difference. The image difference means an index indicating the degree of difference between images indicating the degree of dissimilarity between two images and the degree of dissimilarity. If the image difference between the two images is large, there is a large difference between the two images, which means that the difference is large. On the other hand, if the image difference between the two images is small, it means that the two images are similar. In the case of the same image, the image difference is zero.

  Note that various methods may be applied to calculate the image difference. For example, when the line image acquired by the first line sensor 11 and the line image acquired by the second line sensor 12 are compared, A difference may be obtained by comparing signal values of pixels at the same coordinates, and the absolute value of the difference may be indicated by a sum of all the pixels of the line sensor. In addition, such a comparison is performed not only between line images acquired at the same time but also between line images once stored in the reference image storage unit 120 and line images continuously acquired while the finger 150 is moving. It is also possible to carry out between.

  The image difference calculation unit 130 may include a memory 131 that stores the calculated image difference. In the line movement determination means 300 according to the present embodiment, the reference image difference acquired at a predetermined time such as the initial time is set as a comparison threshold value, and based on this, it is determined whether or not the finger 150 has moved one line or more. Therefore, it is necessary to store the reference image difference as a threshold value. The memory 131 serves to store the reference image difference.

  The partial image data update position setting means 140 is a means for setting a position for determining the next line movement when it is determined that the finger 150 has moved one line or more. Details of the specific calculation contents will be described later.

  Next, with reference to FIG. 11, the contents of calculation control for accurate one-line movement determination by the line movement determination means 300 according to the present embodiment will be described. FIG. 11 is a diagram for explaining the time transition of the line image data acquired by the first line sensor 11 and the second line sensor 12.

  FIG. 11A is a diagram showing line images acquired by the first line sensor 11 and the second line sensor 12 at the first time t = t0. FIG. 11A shows a state in which the first line sensor 11 and the second line sensor 12 are arranged adjacent to each other in parallel, and the finger 150 is swept to the front side. FIG. 11A shows the positional relationship between the finger 150, the first line sensor 11, and the second line sensor 12 at the first time t = t0. When the line image of the fingerprint acquired by the first line sensor 11 at the first time t = t0 is Line A (t0) and the line image of the fingerprint acquired by the second line sensor 12 is Line B (t0), The line image LineB (t0) acquired by the second line sensor 12 is stored in the reference image storage unit 120 as a reference line image. At the same time, the image difference calculation means 130 compares the line image obtained by the first line sensor 11 and the line image obtained by the second line sensor 12 at the first time t = t0, and compares these image differences. α is calculated as a reference image difference. In addition, the reference image difference α at the time t0 may be stored in the memory 131.

  Thereafter, as the finger 150 sweeps and moves as time passes, the line image acquired by the first line sensor 11 and the line image acquired by the second line sensor 12 are LineA (t0) and LineB. The line image at (t0) gradually shifts to the lower side (front side) and changes. Then, when the time when the finger 150 moves the amount corresponding to one of the line sensors 11 and 12 is t1, the line image LineA (t0) at the time t0 is acquired by the second line sensor 12 at the time t1. become.

  FIG. 11B is a diagram showing line images acquired by the first line sensor 11 and the second line sensor 12 at the second time t = t1. In FIG. 11B, when the line image acquired by the second line sensor 12 at the second time t = t1 is LineB (t1), as described above, at time t1, LineA (t0) and LineB ( Since both images of t1) are the same image, they are expressed as in Expression (3).

ここで便宜上、画像差を記号⇔で表現すると、式（３）より次式が導出される。

Here, for convenience, when the image difference is expressed by the symbol ⇔, the following expression is derived from Expression (3).

これは第１の時刻ｔ＝ｔ０における第２のラインセンサ１２で取得及び記憶した基準ライン画像ＬｉｎｅＢ（ｔ０）と、第２のラインセンサ１２で撮像取得するライン画像とを比較して画像差算出手段１３０で両画像の画像差を算出し、これが基準画像差αに達した時刻ｔ１を１ライン移動と判定できることを意味している。つまり、第２のラインセンサ１２で取得されているライン画像と、基準ライン画像ＬｉｎｅＢ（ｔ０）との画像差を取得画像差βとすると、α＝βとなった時刻ｔ１が、指１５０が１ライン分移動したタイミングである。

The image difference calculation is performed by comparing the reference line image LineB (t0) acquired and stored by the second line sensor 12 at the first time t = t0 with the line image acquired and acquired by the second line sensor 12. The means 130 calculates the image difference between the two images, which means that the time t1 when the reference image difference α is reached can be determined as one line movement. That is, if the image difference between the line image acquired by the second line sensor 12 and the reference line image LineB (t0) is an acquired image difference β, the finger 150 is 1 at time t1 when α = β. This is the timing of movement for the line.

  Considering the change in the acquired image difference β from the first time t = t0 until the second time t = t1 is reached, first, the second line sensor 12 at the first time t = t0 is acquired. Since the line image LineB (t0) is naturally the same image in relation to the reference line image LineB (t0), the acquired image difference β (t0) is β (t0) = 0. That is, the following formula (5) is established.

さらに第１の時刻ｔ＝ｔ０から第２の時刻ｔ＝ｔ１への経過において、第１の時刻ｔ＝ｔ０での基準ライン画像ＬｉｎｅＢ（ｔ０）と、ｔ<ｔ１である任意の時刻ｔでの第２のラインセンサ１２で取得されるライン画像ＬｉｎｅＢ（ｔ）との画像差β（ｔ）は、式（６）から式（７）のようになる。

Further, in the passage from the first time t = t0 to the second time t = t1, the reference line image LineB (t0) at the first time t = t0 and the arbitrary time t where t <t1 are satisfied. The image difference β (t) from the line image LineB (t) acquired by the second line sensor 12 is expressed by Expression (6) to Expression (7).

このような関係から、第１の時刻ｔ＝ｔ０における第１のラインセンサ１１のライン画像ＬｉｎｅＡ（ｔ０）と第２のラインセンサ１２のライン画像ＬｉｎｅＢ（ｔ０）同士の基準画像差αに対し、任意の時刻ｔにおける第２のラインセンサ１２で撮像されるライン画像ＬｉｎｅＢ（ｔ）を取得し、取得画像差βがβ≧αとなった時点で、指１５０が１ライン以上移動したと判断することができる。そして、ライン移動の判定を行った段階で、基準ライン画像ＬｉｎｅＢ（ｔ０）と基準画像差αを順次更新するようにすれば、常に更新地点から次の１ライン移動の正確な判定が実施できることを意味する。

From such a relationship, with respect to the reference image difference α between the line image LineA (t0) of the first line sensor 11 and the line image LineB (t0) of the second line sensor 12 at the first time t = t0, A line image LineB (t) captured by the second line sensor 12 at an arbitrary time t is acquired, and it is determined that the finger 150 has moved one line or more when the acquired image difference β becomes β ≧ α. be able to. If the reference line image LineB (t0) and the reference image difference α are sequentially updated at the stage where the line movement is determined, it is possible to always accurately determine the next one line movement from the update point. means.

  In this way, the partial images obtained when moving less than one line are removed each time, and only the partial images obtained when moving more than one line are recorded. It can be removed.

  Such a line movement determination unit 300 may be applied to the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment. The vertical line movement determination is performed by the line movement determination unit 300, and the horizontal movement position when the line movement is determined is executed by the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment. By doing so, it is possible to reduce the calculation processing required for the line movement determination and to quickly perform the calculation of the lateral deviation detection.

  Note that the line movement determination unit 300 may be integrated into the arithmetic processing unit 70 of the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment, or may be applied as a separate and independent apparatus. .

  The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the above-described embodiments, and various modifications and substitutions can be made to the above-described embodiments without departing from the scope of the present invention. Can be added.

It is a functional block diagram of the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 according to the present embodiment. It is explanatory drawing of the estimated pixel data production | generation calculation performed by the estimated pixel data production | generation means 20. FIG. FIG. 2A is a diagram illustrating pixel data of the two-line sensor 10 at the first time t = t0. FIG. 2B is a diagram illustrating pixel data of the two-line sensor 10 at the second time t = t1. It is explanatory drawing of the calculation processing content of the estimation pixel data production | generation means 20 in case the pixel for a finger movement position estimation is six pixels. It is a comparison object figure for demonstrating the arithmetic processing performed with the finger movement position estimation means. FIG. 4A is a diagram illustrating pixel data on the two-line sensor 10 at the first time t = t0. FIG. 4B is a diagram showing pixel data in a state where there is no lateral movement at the second time t = t1. FIG. 4C is a diagram illustrating pixel data in a state where the finger 150 is moved by one line in the vertical direction in the direction of 45 ° to the lower right. It is the figure which showed the example of the pixel data when the finger | toe 150 moved to the horizontal direction below 1 pixel. FIG. 5A is a diagram illustrating pixel data when the finger 150 moves 22.5 ° in the horizontal direction. FIG. 5B is a diagram illustrating pixel data when the finger 150 moves 11.25 ° in the horizontal direction. FIG. 5C is a diagram showing pixel data when the finger 150 moves 33.75 ° in the horizontal direction. FIG. 5D is a diagram showing pixel data when the finger 150 moves 9 ° in the horizontal direction. The lateral movement position information is written in the synthesis memory 90 by the finger movement position estimation means 30. It is a figure showing an example of a sweep mode of a finger. It is the figure which showed an example of the sweep aspect of the finger | toe 150 different from FIG. It is a flowchart of the finger movement position estimation method and image processing method which concern on a present Example. It is the figure which showed schematic structure of the line movement determination means 300 applied to the finger movement position estimation apparatus 100 and the fingerprint image processing apparatus 200 which concern on a present Example. It is explanatory drawing of time transition of the line image data acquired by the 1st line sensor 11 and the 2nd line sensor 12. FIG. FIG. 11A is a diagram showing line images acquired by the first line sensor 11 and the second line sensor 12 at time t0. FIG. 11B is a diagram showing line images acquired by the first line sensor 11 and the second line sensor 12 at time t1.

Explanation of symbols

10 2-line sensor 11, 12 line sensor 20 estimated pixel data generation means 21, 131 memory 30 finger movement position estimation means 40 sweep inclination amount calculation means 45 counting means 50 lateral deviation correction means 60 inclination correction means 70 arithmetic processing means 71 finger movement position Estimation calculation processing means 72 Image correction processing means 80 Support base 90 Synthetic memory 100 Finger movement position estimation device 120 Reference image storage means 130 Image difference calculation means 140 Partial image data update position setting means 150 Finger 160 Calculation control means 200 Fingerprint image processing apparatus

Claims (14)

A finger movement position estimation device having two line sensors extending in the horizontal direction and adjacently arranged in the vertical direction and estimating a movement position and / or movement amount of one finger or less of a finger sweeping on the two line sensors Because
Based on pixel data obtained at a first time by adjacent two rows and two columns of pixels, among the pixels of the two rows and two columns, the movement position of the finger of one pixel in the sweep direction and / or Or estimated pixel data generating means for generating estimated pixel data corresponding to the movement amount;
A finger that estimates the movement position and / or movement amount of the finger by comparing the actual pixel data collected by the one pixel at the second time with the estimated pixel data generated by the estimated pixel data generating means. A finger movement position estimation device. The estimated pixel data generating means generates estimated pixel data corresponding to a position where the finger has moved one line in the vertical direction on the two-line sensor,
The second time is a time when the finger has moved one line in the vertical direction on the two-line sensor,
The finger movement position estimation device according to claim 1, wherein the finger movement position estimation unit estimates the movement position and / or movement amount of the finger in the lateral direction. The estimated pixel data generation means generates a plurality of estimated pixel data at a predetermined lateral movement position obtained by dividing the movement position of one pixel in the horizontal direction and the position moved by the one line in the vertical direction. And
The finger movement position estimation means compares the actual pixel data with estimated pixel data at the predetermined lateral movement position, and among the estimated pixel data at the predetermined lateral movement position, a pixel closest to the actual pixel data The finger movement position estimation apparatus according to claim 2, wherein the predetermined lateral movement position having data is estimated as a lateral movement position of the finger. A synthesis memory that records frame data for one line each time the finger moves on the two-line sensor in the vertical direction by one line;
The finger movement position estimation device according to claim 3, wherein the finger movement position estimation means records lateral movement position information representing the movement position in the horizontal direction together with the frame data in the synthesis memory. Further comprising a counting means for counting the frequency of the lateral movement position information representing the lateral movement position;
Sweep inclination amount calculation means for calculating the amount of sweep inclination of the finger based on the frequency distribution of the lateral movement position information counted by the counting means when the finger sweep is completed. The finger movement position estimation apparatus according to claim 3 or 4. A finger movement position estimation device according to claim 4,
A fingerprint image processing apparatus comprising lateral shift correction means for correcting lateral shift of frame data recorded in a composite memory using lateral movement position information obtained by the finger movement position estimation apparatus. A finger movement position estimation device according to claim 5,
The finger movement position estimation device calculates the amount of sweep inclination of the finger,
A fingerprint image processing apparatus, comprising: a predetermined inclination correcting unit that performs inclination correction of frame data recorded in the composite memory using the sweep inclination amount of the finger. A finger movement position estimation method that estimates a movement position and / or movement amount of one finger or less of a finger that sweeps on two line sensors that extend in the horizontal direction and are adjacently arranged in the vertical direction,
Based on pixel data obtained at a first time by adjacent two rows and two columns of pixels, among the pixels of the two rows and two columns, the movement position of the finger of one pixel in the sweep direction and / or Or an estimated pixel data generation step for generating estimated pixel data corresponding to the movement amount;
Finger movement for comparing the actual pixel data collected by the one pixel at the second time with the estimated pixel data generated by the pixel data generating means and estimating the movement position and / or movement amount of the finger A finger movement position estimation method comprising: a position estimation step. The estimated pixel data generation step generates estimated pixel data corresponding to a position where the finger has moved one line in the vertical direction on the two-line sensor,
The second time is a time when the finger has moved one line in the vertical direction on the two-line sensor,
The finger movement position estimation method according to claim 8, wherein the finger movement position estimation step estimates the movement position and / or movement amount of the finger in the lateral direction. The estimated pixel data generation step generates estimated pixel data at a predetermined lateral movement position obtained by dividing the movement position of one pixel in the horizontal direction and the position moved by the one line in the vertical direction. ,
The finger movement position estimation step compares the actual pixel data with the estimated pixel data at the predetermined lateral movement position, and among the estimated pixel data at the predetermined lateral movement position, a pixel closest to the actual pixel data The finger movement position estimation method according to claim 9, wherein the predetermined lateral movement position having data is estimated as a lateral movement position of the finger. A synthesis memory that records frame data for one line each time the finger moves on the two-line sensor by one line in the vertical direction;
The finger movement position estimation method according to claim 10, wherein lateral movement position information representing the movement position in the horizontal direction is recorded in the synthesis memory together with the frame data. A counting step of counting the frequency of the lateral movement position information indicating the lateral movement position;
A sweep inclination amount calculating step of calculating a sweep inclination amount of the finger based on a frequency distribution of the lateral movement position information counted in the counting step when the finger sweep is completed. The finger movement position estimation method according to claim 10 or 11. Performing the finger movement position estimation method according to claim 11;
A fingerprint image processing method comprising: a lateral displacement correction step for performing lateral displacement correction of frame data recorded in a composite memory using lateral direction movement position information obtained by the finger movement position estimation method. Performing the finger movement position estimation method according to claim 12,
Calculating a finger sweep inclination amount by the finger movement position estimation method;
A fingerprint image processing method, comprising: a predetermined tilt correction step for performing tilt correction of frame data recorded in the composite memory using the sweep tilt amount of the finger.

Images