JP2014106754A - Finger direction specifying system, finger direction specifying method, and program thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technology to be used in sampling a fingerprint, configured to accurately specify a direction (finger direction) indicated by four fingers in a finger image.SOLUTION: A finger direction specifying system includes a finger direction arithmetic processing part 10 which specifies a finger direction indicating a direction of fingers included in finger image data on the basis of input finger image data by image processing. The finger direction arithmetic processing part 10 includes: a dynamic binary image generating part 11 which binary-displays finger sections in the finger image data in a dark state, on the basis of any of a plurality of thresholds set in stages, to obtain dark areas; and a finger direction decision part 14 which decides a directional vector to be obtained by vector-synthesizing length directions of the dark areas of the binary-displayed finger images as a finger direction of the finger image.

Description

  The present invention relates to a fingerprint information storage system that cuts out four finger regions from a finger image, and more particularly, to a finger direction specifying system, a finger direction specifying method, and a program thereof for specifying a finger direction in a finger image.

  Finger photographing systems are used for photographing finger images of a large number of people for immigration and criminal investigations. In recent finger photographing systems, it is required to quickly photograph a finger image of a subject and extract fingerprint information in the finger image. For example, a hand placed on a glass surface like a live scanner is required. A finger photographing device for photographing the image is used.

  Since such a finger photographing device directly reads a finger image optically and stores it as a finger image, each person to be photographed can easily photograph a hand (finger) simply by placing his hand on the glass surface. it can. Further, by using such a finger photographing apparatus, it is possible to effectively reduce the trouble of the photographed person and to shorten the time for photographing.

On the other hand, since the area that can be photographed has widened in the finger photographing device,
Depending on how the hand is placed on the glass surface of the person to be photographed, a part of the palm is reflected in the photographed image.

  By the way, when a finger image is photographed using the finger photographing apparatus as described above, a four-finger region including a fingerprint is cut out from the finger image in order to efficiently extract necessary fingerprint information. It will be necessary.

In order to cut out the four-finger region from the finger image, usually, the finger direction indicating the orientation of the entire finger in the image is determined, and the fingertip side region of the four fingers in the image is specified based on this finger direction.
Here, if the finger direction is misrecognized, there may be a disadvantage that the four-finger region cannot be appropriately cut out.
For this reason, accurately specifying the finger direction can be said to be a very important process in order to perform four-finger segmentation and to appropriately extract fingerprint data from a finger image.

  In general, as a method of determining the finger direction, as shown in FIGS. 12A, 13A, and 14A, the finger image is rotated, and density value accumulation corresponding to each rotation angle is calculated. Then, the angle direction in which the contrast between peaks and valleys in the histogram of accumulated density values is the highest, that is, the direction of the rotation angle at which the angle with the highest degree of separation is set is determined as the finger direction in the finger image. The technique is taken.

  Here, FIGS. 12 (B), 13 (B), and 14 (B) show the density accumulation values in the vertical direction in FIGS. 12 (A), 13 (A), and 14 (A), respectively. It is a histogram.

  In the above method of determining the finger direction based on the accumulated value of density values in the finger image, the finger image having a rotation angle that maximizes the degree of separation is obtained by performing calculations based on the degree of separation of the accumulated density values in these histograms. Each vertical direction (upper edge side of the illustrated drawing sheet) is determined as a finger direction in the finger image.

In FIGS. 12A, 13A, and 14A, a finger image with a rotation angle of 5 degrees (FIG. 13A) is calculated as the maximum separation, and this vertical direction is calculated as follows. It is determined as the finger direction of the finger image.
Here, the vertical direction specifically refers to the case where the XY orthogonal coordinate axes are set on the same plane of the plane frame of FIGS. 12A and 12B as shown in FIG. The direction along the Y axis is shown.

On the other hand, when the finger direction is determined based on the accumulated density value as described above, assuming that a part of the palm is reflected in the captured image, a part of the palm that is reflected in the image processing is used for image processing. There may be a disadvantage that it is erroneously recognized as a part.
In this case, the wrong direction in the finger image may be determined as the finger direction. In such a case, there arises a disadvantage that the four-finger region cannot be accurately cut out.

  For example, in the finger image shown in FIG. 15 (A), a part of the palm is reflected. Here, when the density value histogram is generated based on the accumulated density value in the finger image, the difference between the high and low accumulated density values in the density value histogram becomes small as shown in FIG. The degree of separation between the density values corresponding to each of the four fingers becomes ambiguous.

  Further, when the accumulated density value is calculated for a finger image obtained by rotating the finger image of FIG. 15A to the right, as shown in FIG. 16A, an incorrect finger direction is determined. That happens.

  For example, in the case of FIG. 15A, the density value histogram generated by the image processing based on FIG. 15A is shown in FIG. As shown in FIG. 15B, the palm region reflected in the finger image is erroneously recognized as one of the four fingers.

  Furthermore, when four-finger clipping is performed based on the wrong finger direction determined here (FIG. 16A), the four fingers in the image are not recognized correctly, and as shown in FIG. The long side portion of the recognition frame is extremely inclined, or the finger recognition frame itself is concentrated at the offset position, and erroneous finger recognition is performed.

  As described above, if the finger direction in the finger image is erroneously determined, the four fingers cannot be appropriately cut out and incorrect finger recognition is performed, so that the fingerprint data cannot be properly extracted from the finger image. The inconvenience arises.

  As a related technique related to the above-described series of finger image recognition, there is known a system for increasing the fingerprint collation rate by performing accurate image rotation correction when the fingerprint image is inclined (Patent Document 1). ).

  Similarly, as a related technique, there is known a fingerprint collation apparatus that calculates a tilt angle of a fingerprint image for fingerprint collation while suppressing the influence of fingerprint sensor variations and fingerprint image quality (Patent Document 2).

JP 2001-76145 A JP 2008-217090 A

  However, in the related techniques described in Patent Documents 1 and 2, there is no assumption that a palm is in the photographed finger image. For this reason, in these related technologies, there is a problem that an incorrect finger direction is calculated in the finger image due to the influence of the captured palm region.

  When the wrong finger direction is calculated, as described above, the four-finger region cannot be appropriately cut out from the finger image, so that there is a disadvantage that wrong finger recognition is performed on the finger image.

[Object of invention]
The present invention provides a finger direction specifying system, a finger direction specifying method, and a program thereof that can improve the inconvenience of the related technology and can appropriately specify the finger direction indicating the direction in which all four fingers are facing in the finger image. Its purpose is to provide.

In order to achieve the above object, a finger direction specifying system according to the present invention uses a finger direction calculation to specify a finger direction indicating a finger direction included in the finger image data based on input finger image data by image processing. A finger direction identification system including a processing unit,
The finger direction calculation processing unit,
An appropriate state obtained by binarizing each finger part of the finger image into a dark color state based on one specific threshold value among a plurality of threshold values set in stages for the finger image data And a direction vector obtained by vector-combining the length directions of the dark color areas of the respective binarized finger images to obtain a finger direction of the finger image. And a finger direction deciding unit for deciding.

In order to achieve the above object, the finger direction specifying method according to the present invention includes a finger direction calculation process for specifying a finger direction indicating a finger direction included in the finger image by image processing based on input finger image data. In the finger direction identification system with a part,
Smooth the finger image data input externally,
The smoothed finger image data is binarized into a dark state for each finger portion of the finger image based on one specific threshold value among a plurality of threshold values set in stages. To create a dark area in the proper state
The binarized display of the finger image of each finger image is synthesized with a vector in the length direction, and the direction vector obtained by the vector composition is determined as the finger direction of the finger image. To do.

In order to achieve the above object, a finger direction specifying program according to the present invention is a finger direction calculation that specifies, by image processing, a finger direction indicating a finger direction included in the finger image data based on input finger image data. In a finger direction identification system with a processing unit,
A smoothing function that smoothes finger image data that is input externally,
The smoothed finger image data is binarized into a dark state for each finger portion of the finger image based on one specific threshold value among a plurality of threshold values set in stages. Binarization processing function to generate dark areas in the proper state
And a finger direction determination processing function for determining a vector of the length direction of the dark color region of each binarized finger image and determining the direction vector obtained by the vector combination as the finger direction of the finger image. Provided,
A finger direction specifying program characterized in that each of these processing functions is realized by a computer provided in the finger direction calculation processing unit.

  Since the present invention is configured as described above, according to this, the finger direction indicating the direction of the finger included in the finger image can be accurately specified without being affected by the palm area reflected in the finger image. It is possible to provide an excellent finger direction specifying system, a finger direction specifying method, and a program thereof that are not included in the related technology.

It is a schematic block diagram which shows the principal part of 1st Embodiment in the finger direction identification system concerning this invention. It is a schematic block diagram which shows 1st Embodiment containing the principal part disclosed in FIG. It is a schematic block diagram which shows the principal part of 2nd Embodiment in the finger direction identification system concerning this invention. It is a schematic block diagram which shows 2nd Embodiment containing the principal part disclosed in FIG. It is a flowchart which shows operation | movement of the principal part in 2nd Embodiment disclosed in FIG. FIG. 6 is a diagram illustrating an example of an input image (finger image) in the second embodiment disclosed in FIG. 3, and FIG. 6A is an explanatory diagram illustrating a finger image in which a part of a palm of four fingers is reflected; FIG. 6B is an explanatory diagram showing a finger image when the smoothing process of FIG. FIG. 7B shows four examples when the smoothed finger image shown in FIG. 6B is sequentially binarized at a certain threshold interval, and FIG. 7A is dark due to a preset threshold value. FIG. 7B shows an example of a binarized image in which pixels are output as a dark color area, and FIG. 7B shows an example of a binarized image when the threshold value is lowered by one step compared to the binarized image of FIG. FIG. 7C shows an example of a binarized image when the threshold value is lowered by one step compared to the binarized image of FIG. 7B, and FIG. 7D shows an example of FIG. It is explanatory drawing which shows an example of the binarized image at the time of lowering | hanging a threshold value by one step rather than the binarized image. It is explanatory drawing which shows the area direction of each area | region derived | led-out in the binarized image in the finger image processed in 2nd Embodiment disclosed in FIG. 3, and the line segment of a finger direction. It is explanatory drawing which shows an example of the recognition result at the time of performing the recognition process of a fingertip area | region for the finger image of FIG. 6 (A) based on the finger | toe direction derived | led-out by 2nd Embodiment disclosed in FIG. It is explanatory drawing which shows an example of the finger image which made object the plane 4 finger made into the object of processing in 2nd Embodiment disclosed in FIG. It is explanatory drawing which shows an example of the result of the finger recognition process performed by making the finger image of FIG. 10 into object. FIGS. 12A and 12B are diagrams illustrating an example of a plane four-finger image for only four plane fingers. FIG. 12A is an explanatory diagram illustrating an example of a plane four-finger image with a rotation angle of 0 degree, and FIG. FIG. 12A is an explanatory diagram showing an example of a histogram of vertical density accumulation values in the image of A), and FIG. 12C is an explanatory diagram showing an XY coordinate system of the finger image disclosed in FIG. FIGS. 13A and 13B are diagrams showing another example of a four-finger image targeting only four flat fingers. FIG. 13A is an explanatory diagram showing an example of a four-finger image having a rotation angle of 5 degrees, and FIG. It is explanatory drawing which shows the example which made the density | concentration accumulated value of the vertical direction in the image of 13 (A) into the histogram. FIG. 14A is a diagram showing still another example of a four-finger image targeting only four fingers, FIG. 14A is an explanatory diagram showing an example of a four-finger image having a rotation angle of 45 degrees, and FIG. It is explanatory drawing which shows the example at the time of carrying out the histogram of the density | concentration accumulated value of the orthogonal | vertical direction in the image of FIG. FIG. 15A is a diagram illustrating an example of a flat four-finger image in which a part of a palm is reflected in a finger image, and FIG. 15A is an explanatory diagram illustrating an example of a flat four-finger image with a rotation angle of 0 degrees, and FIG. () Is an explanatory diagram showing an example in which the accumulated density value in the vertical direction in the image of FIG. FIG. 16A is a diagram showing still another example of a planar four-finger image in which a part of the palm is reflected in the finger image, and FIG. 16A is an explanatory diagram showing an example of a planar four-finger image having a rotation angle of 45 degrees to the right; FIG. 16B is an explanatory diagram showing an example in which the accumulated density values in the vertical direction in the image of FIG. It is explanatory drawing which shows an example of the incorrect finger recognition result performed based on the plane 4 finger image in which a part of palm shown by FIG. 15 (A) is reflected in the finger image.

  First, in the first embodiment, the basic configuration content of the finger direction system according to the present invention will be described. Next, in the second embodiment, the specific configuration content of the finger direction determination system according to the present invention will be described. explain.

[First Embodiment]
First, a first embodiment of a finger direction determination system according to the present invention will be described with reference to FIGS.

  In the first embodiment, the finger direction determination system 100 specifies a finger direction indicating the direction of a finger included in the finger image data based on externally input finger image data by image processing. It has. The finger direction calculation processing unit 10 is configured with a CPU as a base (see FIG. 2).

  Further, the finger direction determination system 100 includes a first storage unit 22 that stores a program for realizing the finger direction calculation processing unit 10 on the CPU, and operations of components of the finger direction calculation processing unit 10. A second storage unit 23 that stores specific data necessary for realizing the function and a result of the operation, an output display unit 24 that displays and outputs the operation state of the finger direction calculation processing unit 10 to the outside, and a finger direction calculation process And a processing data output unit 25 for outputting processing data from the unit 10 to the outside.

  Here, the reference numeral 24A indicates a display image storage unit that stores a display image as necessary. Reference numeral 10A denotes an input circuit unit for inputting image data and other necessary information including an operation command from the outside.

  Among these, as shown in FIG. 1, the above-described finger direction calculation processing unit 10 converts the above-described finger image data input from the outside into a finger image based on a plurality of threshold values set in stages. A dynamic binarized image generation unit 11 that obtains a dark color area (hereinafter referred to as “black area”) by sequentially binarizing into a dark color state (black state when binarizing with black and white) is provided.

  Further, the finger direction calculation processing unit 10 targets a black region of each finger image binarized and displayed by the dynamic binarized image generation unit 11 and a direction obtained by vector synthesis of the length directions. A finger direction determination unit 14 is provided for determining a vector as the finger direction of the finger image.

That is, since the binarization process specifies the length direction of the black area of each finger part, the finger direction determination unit vector-synthesizes the direction along the longest area of each finger part, thereby It is possible to effectively specify the finger direction according to the finger image data.
Here, the input grayscale finger image may be smoothed in advance by a smoothing processing unit.

The above-described dynamic binarized image generation unit 11 sets a threshold value stepwise from the darkest color depth to the brighter color depth for a finger image input externally in order to realize an uplifted function. The binarized image processing unit 12 is provided.
The binarized image processing unit 12 outputs a binarized image with a pixel region darker than the threshold value as a black region and a bright pixel region as a white region with reference to preset threshold values.

  Further, the dynamic binarized image generation unit 11 has each black region corresponding to a set threshold among different black regions in the binarized image processed by the binarized image processing unit 12. When expanded, the image forming apparatus includes an adjacent region specifying unit 13 that specifies a binarized region that is connected to a black region that is adjacent to the first as an adjacent region.

  And this adjacent area specific | specification part 13 makes the state which stopped the display state of each black area which is the said adjacent area to the display state of the stage immediately before connecting with another black area, and makes this into a binarized image. It has a function to acquire and hold.

Specifically, the above-described finger direction determination unit 14 performs, for each black region in the binarized image related to the adjacent region specified by the adjacent region specifying unit 13 of the dynamic binary image generation unit 11, It has a function of setting the direction of the black pixel region having the longest distance from the centroid pixel as the region direction, starting from the centroid pixel in each black region.
Then, as described above, the finger direction determination unit 14 determines a direction vector obtained by combining all the region directions in the binarized image as the finger direction of the input finger image.

Next, the operation of the first embodiment will be described.
First, the dynamic binarized image generation unit 11 operates on the image data of the input four-finger image, and the input four-finger image is stepped from the darkest color depth to the brighter color depth in advance. A binarized image having a pixel area darker than the threshold value as a black area is generated on the basis of each threshold value set to (binarized image generating step).

Here, the binarized image generation step is specifically executed by the binarized image processing unit 12 of the dynamic binarized image generation unit 11. In this case, a pixel area brighter than the threshold value is set as a white area.
Prior to the execution of the binarized image generation step, the smoothing processing unit is caused to function so as to perform smoothing processing on the input finger image (smoothing processing step). The binarization process may be performed on an image (smooth image).

Next, the dynamic binarized image generation unit 11 sequentially expands the areas corresponding to the threshold values set in stages among a plurality of different black areas in the generated binarized image. However, in the process, if there is a state where it is connected to an adjacent black area, this is specified as an adjacent area, and the function is provided to maintain the adjacent black area immediately before the connection as an appropriate area. ing. Specifically, the adjacent region specifying unit 14 of the dynamic binarized image generating unit 11 executes the specification of the adjacent region and the acquisition and holding of the appropriate region.
That is, when the adjacent area is specified, the adjacent area specifying unit 14 performs a process of keeping the display state of each black area, which is the adjacent area, in a display state (appropriate state) immediately before being connected to another black area. (Adjacent region specifying step).

  Subsequently, for the binarized image in an appropriate state, the finger direction determination unit 14 starts from the barycentric pixel in each black area and reaches the black pixel with the longest distance from the barycentric pixel in each black area. The direction is derived as the region direction. Also, the finger direction determination unit 14 determines a direction vector generated by combining all the region directions in the binarized image as the finger direction of the finger image.

Here, regarding the smoothing processing step, the binarized image output step, the adjacent region specifying step, and the finger direction determining step, the execution contents may be programmed and executed by a computer.
In addition, the program may be recorded on a non-transitory storage medium such as a DVD, a CD, or a flash memory. In this case, the program is read from the recording medium by a computer and executed.

  As described above, the finger direction calculation system 100 can perform the finger image even when the finger direction calculation processing unit 10 as the main part functions effectively and a part of the palm is reflected in the finger image as described above. It is possible to appropriately specify the finger direction at.

[Second Embodiment]
Next, a second embodiment of the present invention will be described with reference to FIGS.
Here, the same reference numerals are used for the same constituent members as those of the first embodiment described above.
In the second embodiment, the finger direction determination system 300, as in the case of the first embodiment described above, is based on the finger image data input from the outside, and the finger direction indicating the direction of the finger included in the finger image data. Is provided with a finger direction calculation processing unit 30 that identifies the image by image processing. The finger direction calculation processing unit 30 is configured with a CPU as a base (see FIG. 4).

  Further, the finger direction determination system 300 includes a first storage unit 42 that stores a program for realizing the finger direction calculation processing unit 30 on the CPU, and operations of respective components of the finger direction calculation processing unit 30. And a second storage unit 43 that stores specific data necessary for realizing the function and a result of the operation.

  As shown in FIG. 3, the finger direction determination system 300 further includes an input circuit unit 30 </ b> A that acquires a gray scale (for example, 255 gradation) finger image and a smoothing process that performs a smoothing process on the finger image. A binarization processing unit 30B, a dynamic binarization processing unit 31 that sequentially sets threshold values for the smoothed finger image and sequentially generates a binarized image based on each threshold value; The region direction indicating the direction of each region (black region which is a finger portion) of one specific binarized image in the output binarized image is obtained, and the finger of the entire finger image in the region direction is obtained. And a finger direction determining unit 34 for determining the direction.

  Among these, the dynamic binarization processing unit 31 and the finger direction determination unit 34 respectively function in the same manner as the dynamic binarized image generation unit 11 and the finger direction determination unit 14 in the first embodiment described above. It has become a thing.

The input circuit unit 30A adds an image edge having a predetermined size to the input planar four-finger image (see FIG. 6A: hereinafter referred to as “finger image”), thereby binarizing the image. The processing range setting unit 30a for setting the conversion processing range is provided.
In other words, the processing range setting unit 30a has a function of setting a region range to be smoothed, and has a function corresponding to the trimming means.

  The smoothing processing unit 30B described above performs a smoothing process on the trimmed finger image (FIG. 6B). Here, the smoothing process by the smoothing processing unit 30B indicates a process of blurring the finger image by, for example, taking the average of the colors of eight points near the designated pixel.

  Further, the smoothed image (smoothed image) shown in FIG. 6B is a grayscale image of 255 gradations, in which the color depth of white pixels is 0 and the color depth of black pixels is 255. And

The dynamic binarized image generation unit 31 described above sequentially binarizes and displays the smoothed image data in different dark color areas (black areas) based on a plurality of threshold values set stepwise. A binarized image processing unit 32 is provided.

  Further, the dynamic binarized image generation unit 31 is expanded corresponding to the threshold value among the black areas in the binarized image, so that the black area that is connected to the black area in the close position is connected. An adjacent area specifying unit 33 that monitors and specifies the area is provided. In this case, the adjacent area specifying unit 33 specifies a black area (hereinafter referred to as “area”) connected to the adjacent black area as the adjacent area.

At the same time, the adjacent area specifying unit 33 includes an appropriate state holding function 33A which is a processing function (acquires and holds) the state immediately before the display state of each area that is an adjacent area is connected to another area as an appropriate state. ing.
As a result, the adjacent area maintains the shape and size of the area without expanding the area as it was immediately before the display area (appropriate state) connected to the other area, and other areas correspond to threshold values. To be expanded.

  In addition, the above-described binarized image processing unit 32, specifically, the threshold value (T × n) is stepped at regular intervals from the darkest color depth to the brighter color depth with respect to the smoothed finger image. ) And a binarized image having a pixel area darker than the set threshold value as a black area and a bright pixel area as a white area.

  When the threshold value is set for the smoothed finger image, the binarized image processing unit 32 sets the threshold level at equal intervals from the darkest color depth to the preset color depth. A stepwise threshold setting function 32A for setting is provided.

  Specifically, the binarized image processing unit 32 sets 240 as the threshold value (T × 1). A binarized image generated based on this threshold is shown in FIG. In FIG. 7A, pixels with a color depth of 240 to 255 in the smoothed image are output as black areas. Further, as the number of black areas on the four fingers, 11 black areas are output and displayed in FIG. 7A.

Next, the binarized image processing unit 32 sets 230 as a threshold value (T × 2), and a binarized image generated based on the threshold value is shown in FIG.
In FIG. 7B, pixels with a color depth of 230 to 255 in the smoothed image are output as black areas.

  Here, it is shown that the 11 black areas in the binarized image (FIG. 7B) are expanded as compared with FIG. 7A.

  FIG. 7C shows a binarized image generated by setting 220 as the threshold value (T × 3) and based on this threshold value. In FIG. 7C, pixels with a color depth of 230 to 255 in the smoothed image are output as black areas.

Further, FIG. 7C shows a state immediately before the region 10 and the region 11 in FIG. 7A are connected. In this case, the appropriate state holding function 33A of the adjacent area specifying unit 33 functions to keep the display states of the area 10 and the area 11 as the adjacent areas in this state (normal state).
As a result, the regions 10 and 11 indicating the finger regions are in a state immediately before being connected, and the shape and size are maintained without being expanded even when a threshold value is newly set (FIG. 5 ( D)).

  Also, the finger direction determination unit 34 described above determines the center-of-gravity pixel in each region in the binarized image, as in the case of the first embodiment described above, and the distance from the center-of-gravity pixel among the black pixels in each region. The black pixel with the longest is specified. The finger direction determination unit 34 includes a region direction specifying function 34A that sets the length direction of the line segment from the centroid pixel to the black pixel having the longest distance from the centroid pixel as the region direction. .

  Further, the finger direction determination unit 34 sets the region direction corresponding to each of all the regions in the binarized image. The finger direction determination unit 34 includes a finger direction determination processing function 34B that determines a direction vector obtained by combining all the region directions as a finger direction corresponding to the finger image.

Here, FIG. 8 shows the region direction of each region (regions 1 to 11) specified by the finger direction determination unit 34 and the finger direction corresponding to the binarized image.
In FIG. 8, finger numbers 1 to 4 derived based on the specified finger direction are assigned.

The finger direction determination unit 34 compares the difference between the finger direction determined by combining the region directions and the region direction, excludes the region direction having the largest difference, and then combines the region directions again. By performing this, a composite direction update function 34C for updating the finger direction is provided.
When such a method is employed, the accuracy of the finger direction once specified can be further increased.
Other configurations and operations thereof are the same as those of the first embodiment described above.

  With the configuration described above, in the second embodiment, it is possible to derive an appropriate finger direction from the finger image even when a part of the palm area is reflected in the finger image.

  For this reason, when the four-finger segmentation is performed based on the finger direction determined according to the second embodiment, the four-finger region can be identified and segmented more reliably, and furthermore, the fingerprint can be identified from the finger image. It is possible to effectively suppress the occurrence of erroneous recognition of the finger position area in the data collection system.

(Operation of Second Embodiment)

Next, the operation of the finger direction determination system in the second embodiment will be described based on the flowchart of FIG.
First, the basic operation of the direction determination system 300 according to the second embodiment will be described, and then specific operation contents will be described sequentially.

First, the smoothing processing unit 30B smoothes the finger image data input from the outside.
Thereafter, the smoothed finger image data is operated based on one specific threshold value among a plurality of threshold values set stepwise, and the dynamic binarized image generation unit 31 operates to Each finger part of the image is binarized and displayed in a dark state to generate a dark region (a dark region in the case of black and white display).

  Thereafter, the finger direction determination unit 34 performs vector synthesis on the length direction of the binarized dark color area of each finger image, and the direction vector obtained by the vector synthesis is used as the finger direction of the finger image. decide.

Here, in the step of binarizing and displaying each finger portion of the finger image in the dark color area to generate the dark color area, the above-described binarized display of the finger image data is set in advance in steps. The entire finger images are sequentially binarized and displayed in different dark areas based on the plurality of threshold values.
After that, the state of the binarized display so that an appropriate state immediately before any one of the dark color areas to be binarized and connected to another adjacent dark color area is obtained. To monitor. When the appropriate state is obtained, the binarized finger image of the appropriate state is specified and held.

This will be described more specifically below.
First, when the finger image data shown in FIG. 6A is input to the finger direction determination system 300 via the input circuit unit 30A, the processing range setting unit (corresponding to the trimming processing unit) 30a of the input circuit unit 30A Functions, and sets a rectangular area (image edge) of a certain size for the image data as shown by the frame edge in FIG. 6A (FIG. 5: step S101 / image edge adding step).

Next, in the smoothing processing unit 12, the smoothing processing unit 30B functions on the finger image to which the image edge is added, and smoothing processing is performed (FIG. 5: step S102 / smoothing processing step). An example of the smoothed finger image is shown in FIG.
Here, it is assumed that the smoothed image (smoothed image) is a grayscale image having 255 gradations. The color depth values in the second embodiment are assumed to be 255 for black and 0 for white.

  Subsequently, the binarized image processing unit 32 of the dynamic binarized image generation unit 31 performs a threshold stepwise at regular intervals from the darkest color depth to the brighter color depth with respect to the smoothed finger image. (T × n) is set. Then, a binarized image in which the pixel area darker than the set threshold is a black area and the pixel area brighter than the threshold is a white area is output (FIG. 5: Step S103 / binarized image processing step).

Subsequently, a plurality of threshold values are sequentially applied to perform binarization processing, and whether or not a proper state of the above-described binarized image is generated is determined and acquired as follows (FIG. 5: Step S104 / Appropriate state acquisition process).
First, the binarized image processing unit 32 sets 240 as a threshold value (T × 1), and a binarized image generated based on the threshold value is shown in FIG.
In FIG. 7A, pixels with a color depth of 240 to 255 in the smoothed image are output as black areas. Here, 11 black areas (areas 1 to 11) are output as shown.

  Next, the binarized image processing unit 32 sets 230 as a threshold value (T × 2), and a binarized image generated based on the threshold value is shown in FIG. In FIG. 7B, pixels with a color depth of 230 to 255 in the smoothed image are output and displayed as black areas.

  7B shows that each of the 11 black areas in the binarized image is expanded as compared with FIG. 7A.

  Next, the adjacent region specifying unit 33 of the dynamic binarized image generation unit 31 is expanded corresponding to the threshold value set stepwise in the black region in the binarized image, An area that is connected to the area is identified as an adjacent area.

Here, the adjacent area specifying unit 33 performs a process of keeping the display state of each area, which is an adjacent area, in a state immediately before being connected to another area (normal state) (FIG. 5: Step S105 / adjacent area expansion suppressing step). .
As a result, the shape and size of the area are maintained without expanding the adjacent area in the area (normal state) which is the display state immediately before being connected to another area. Note that the areas other than the adjacent areas are expanded corresponding to the threshold value set in stages.

  Here, 220 is set as the threshold value (T × 3), and a binarized image generated based on this threshold value is shown in FIG. In FIG. 7C, pixels with a color depth of 230 to 255 in the smoothed image are output as black areas.

Further, FIG. 7C shows that the region 10 and the region 11 are just before being connected. The adjacent area specifying unit 33 keeps the display states of the area 10 and the area 11 as the adjacent areas in this state (normal state).
As a result, as shown in FIG. 7D, the regions 10 and 11 are maintained in the state immediately before they are connected to each other even when a threshold value is newly set, and their shapes and sizes are not expanded. Is held (FIG. 5: Step S105).

  Note that the binarized image processing unit 32 sets threshold values step by step up to a preset color depth for other regions (regions 1 to 9 other than adjacent regions), and corresponds to these threshold values. Perform binarized output.

Next, as described above, the finger direction determination unit 34 functions to determine the finger direction in the finger image as follows.
First, the finger direction determination unit 34 specifies each barycentric pixel for each region (black region) of the binarized image generated by the dynamic binarized image generation unit 31.

  Subsequently, the finger direction determination unit 34 identifies a pixel (distant edge pixel) having the longest distance from each centroid pixel in each region, and sets the direction from the centroid pixel to this far edge pixel as the region direction in that region. Define (Figure 8). Thereby, the finger direction determination unit 34 determines the region direction for each region (region direction determination processing function).

In addition, when determining the direction from the centroid pixel to the far edge pixel, the finger direction determination unit 34 sets a pixel within a range of 180 degrees on the upper side of the finger image as a far edge pixel candidate pixel based on the position of the centroid pixel.
As a result, each region direction is defined as a vector directed upward in the finger image.

  Subsequently, the finger direction determination unit 34 synthesizes the region directions set corresponding to all the regions set in the binarized image (FIG. 5 / regions 1 to 11). Thereby, the generated direction (combined vector) is determined as the finger direction in the binarized image (ie, finger image) (FIG. 5: Step S106 / finger direction determining step).

Here, the operation content of each step in the second embodiment described above may be configured such that the execution content is programmed and executed by a computer.
In addition, the program may be recorded on a non-transitory storage medium such as a DVD, a CD, or a flash memory. In this case, the program is read from the recording medium by a computer and executed.

  In each of the first and second embodiments, the case where the density state (1/0 processing) is set to black and white display is exemplified in the binarization processing. Color may be used.

As described above, in the finger direction determination system 300 according to the second embodiment, as in the case of the first embodiment described above, the finger direction calculation processing unit 30 that is a main part functions effectively, and the palm image is included in the finger image. Even when a part of the image is reflected, it is possible to appropriately specify the finger direction in the finger image.
Furthermore, in the second embodiment, the image smoothing processing by the smoothing processing unit 30B can obtain a stable binarized image from which noise has been eliminated even during the binarization processing. In the synthesis, the vector synthesis can be performed with high accuracy, and in this respect, the reliability in determining the finger direction can be greatly improved. Therefore, the finger direction in the finger image can be accurately specified. It has the advantage that it is possible.

  As a result, the four-finger cutout from the finger image can be more reliably performed based on the finger direction. Therefore, in the system for collecting fingerprint data, as shown in FIG. It is possible to more reliably detect the fingertip area of each of the four fingers.

About the above-mentioned embodiment, it is as follows when the summary of the novel technical content is put together.
In addition, although one part or all part of the said embodiment is put together as follows as a novel technique, this invention is not necessarily limited to this.

[Appendix 1]
A finger direction specifying system including a finger direction calculation processing unit that specifies, by image processing, a finger direction indicating a finger direction included in the finger image data based on input finger image data,
The finger direction calculation processing unit,
An appropriate state obtained by binarizing each finger part of the finger image into a dark color state based on one specific threshold value among a plurality of threshold values set in stages for the finger image data A dynamic binarized image generation unit for obtaining a dark color area of
And a finger direction determining unit that determines a direction vector obtained by vector synthesis of the length directions of the dark color areas of each binarized image of the fingers as the finger direction of the finger image. Finger direction identification system.

[Appendix 2]
In the finger direction identification system according to attachment 1,
The dynamic binarized image generation unit
A binarized image processing unit that sequentially binarizes the finger image data into different dark color regions based on a plurality of threshold values set in stages, and each dark color region that is binarized and displayed An adjacent region specifying unit that monitors whether or not a proper state that is a binarized state immediately before any one of the dark region is connected to another adjacent dark region,
The finger direction specifying system, wherein the adjacent region specifying unit has a proper state holding function for acquiring and holding the binarized finger image when the proper state is obtained.

[Appendix 3]
In the finger direction identification system according to attachment 2,
The binarized image processing unit has a stepwise threshold setting function for setting the threshold stepwise from the darkest color depth to a preset color depth when setting the threshold value. Finger direction identification system characterized by.

[Appendix 4]
In the finger direction identification system according to attachment 2,
A processing range setting unit for adding an image frame of a predetermined size to finger image data input externally is provided in the data input stage for the finger image of the dynamic binarized image generation unit. Finger direction identification system characterized by.

[Appendix 5]
In the finger direction identification system according to attachment 2,
A finger direction specifying system, wherein a smoothing processing unit that smoothes externally input finger image data is provided at a data input stage related to the finger image of the dynamic binary image generation unit.

[Appendix 6]
In the finger direction identification system according to any one of appendices 1 to 5,
The finger direction determination unit
An area direction specifying function in which the length direction of the line segment from the centroid pixel of each dark color area as a starting point to the dark color pixel having the longest distance from the centroid pixel is the area direction; A finger direction determination system comprising: a finger direction determination processing function for determining a direction vector formed by vector composition of region directions as a finger direction of the finger image.

[Appendix 7]
In the finger direction identification system according to any one of appendices 1 to 5,
The finger direction determination unit
While having a region direction specifying function in which the length direction of the line segment from the centroid pixel of each of the dark color regions generated to the dark color pixel having the longest distance from the centroid pixel is the region direction,
The finger direction specifying system, wherein the region direction in each dark color region is determined within a range of 180 degrees left and right on the extension side of each finger image.

[Appendix 8]
In the finger direction identification system according to any one of appendices 1 to 5,
The finger direction determination unit
While having a finger direction determination processing function for determining a direction vector formed by vector composition of the region directions of each line segment as a finger direction of the finger image,
The region direction of the remaining line segments is vector-synthesized by excluding the line segment in the region direction where the difference between the finger direction specified by the finger direction determination processing function and the region direction of the original line segment is the largest. A finger direction specifying system comprising a composite direction updating function for updating a finger direction.

[Supplementary Note 9] (Invention of Method / Corresponding to Supplementary Note 1)
In a finger direction identification system including a finger direction calculation processing unit that identifies a finger direction indicating a direction of a finger included in the finger image based on input finger image data by image processing,
Smooth the finger image data input externally,
Each finger portion of the finger image is binarized and displayed in a dark color state based on one specific threshold value among a plurality of threshold values set in a stepwise manner on the smoothed finger image data. To create a dark area in the proper state
While vector-combining the length direction of the dark color area of each binarized displayed finger image, determining the direction vector obtained by this vector composition as the finger direction of the finger image,
A finger direction specifying method characterized in that the finger direction calculation processing unit sequentially executes these steps.

[Appendix 10] (Appendix 2)
In the finger direction identification method according to attachment 9,
In the step of binarizing and displaying each finger part of the finger image in a dark color area,
In the binarized display of the finger image data, the entire finger image is sequentially binarized and displayed in different dark areas based on a plurality of threshold values set stepwise in advance,
The state of the binarized display is monitored so that an appropriate state immediately before any one of the dark color areas to be binarized displayed is connected to another adjacent dark color area is obtained. And
A finger direction specifying method characterized by specifying and holding a binarized finger image of the appropriate state when the appropriate state is obtained.

[Appendix 11] (Appendix 6)
In the finger direction specifying method according to appendix 9 or 10,
In the step of determining the finger direction of the finger image,
The area direction is the length direction of the line segment from the centroid pixel of each generated dark color area to the dark color pixel having the longest distance from the centroid pixel,
Thereafter, the finger direction specifying method is characterized in that a direction vector formed by vector synthesis of the region directions of each line segment is determined as the finger direction of the finger image.

[Appendix 12] (Appendix 8)
In the finger direction specifying method according to appendix 9 or 10,
In the finger direction determination step,
After determining the direction vector formed by vector composition of the area direction of each line segment as the finger direction of the finger image,
Excludes the line segment in the area direction where the difference between the determined finger direction and the area direction of each original line segment is the largest, and vector-combines the remaining line segment area directions to obtain the normal finger direction. A finger direction specifying method characterized by that.

[Appendix 13] (Program invention / Appendix 9)
In a finger direction specifying system including a finger direction calculation processing unit that specifies a finger direction indicating a finger direction included in the finger image data based on input finger image data by image processing,
A smoothing function that smoothes finger image data that is input externally,
The smoothed finger image data is binarized into a dark state for each finger portion of the finger image based on one specific threshold value among a plurality of threshold values set in stages. Binarization processing function to generate dark areas in the proper state
And a finger direction determination processing function for determining a vector of the length direction of the dark color region of each binarized finger image and determining the direction vector obtained by the vector combination as the finger direction of the finger image. Provided,
A finger direction specifying program characterized in that each of these processing functions is realized by a computer provided in the finger direction calculation processing unit.

[Appendix 14] (Appendix 10)
In the finger direction identification program according to attachment 13,
The binarization processing function
In the binarization processing of the finger image data, a binarization sequential processing function that sequentially binarizes the entire finger image into different dark areas for each of a plurality of threshold values set stepwise in advance,
Each of the fingers to be binarized so that an appropriate state immediately before any one of the dark color areas to be binarized is connected to another adjacent dark color area is obtained. Processing status monitoring function that monitors the status of the entire image,
And a binarized image specifying and holding function for specifying and holding a binarized finger image in the proper state when the appropriate state of each dark color region is obtained,
A finger direction specifying program characterized in that each of these functions is realized by the computer.

[Appendix 15] (Appendix 11)
In the finger direction identification program according to appendix 13 or 14,
The finger direction determination processing function
An area direction specifying processing function in which the length direction of the line segment from the centroid pixel of each of the dark color areas as a starting point to the dark color pixel having the longest distance from the centroid pixel is the area direction;
A direction vector determination processing function for determining a direction vector formed by vector composition of the specified region directions of each line segment as a finger direction of the finger image;
A finger direction specifying program characterized in that each of these functions is realized by the computer.

[Appendix 16] (Appendix 12)
In the finger direction identification program according to appendix 13 or 14,
The finger direction determination processing function
After determining the direction vector formed by vector composition of the area directions of each line segment as the finger direction of the finger image, the difference between the determined finger direction and the area direction of each original line segment is calculated, and the difference Specific line segment exclusion processing function that excludes the line segment that extends in the direction of the largest area,
And a finger direction specifying processing function for specifying a new finger direction obtained by vector synthesis of the region directions of the remaining line segments as a true finger direction instead of the determined finger direction,
And a finger direction specifying program characterized in that each of these functions is realized by the computer.

  The present invention can be usefully applied to a fingerprint information storage system that cuts out four fingers in a finger image and stores the four head regions of each of the four fingers as image information.

10, 30 Finger direction calculation processing unit 10A, 30A Input circuit unit 11, 31 Dynamic binarized image generation unit 12, 32 Binarized image processing unit 13, 33 Adjacent region specifying unit 14, 34 Finger direction determining unit 30a Processing Range setting unit 30B Smoothing processing unit 32A Stepwise threshold setting function 33A Adjacent region specification holding function 34A Region direction specification function 34B Finger direction determination processing function 34C Composite direction update function 100, 300 Finger direction determination system

Claims (16)

A finger direction specifying system including a finger direction calculation processing unit that specifies, by image processing, a finger direction indicating a finger direction included in the finger image data based on input finger image data,
The finger direction calculation processing unit,
An appropriate state obtained by binarizing each finger part of the finger image into a dark color state based on one specific threshold value among a plurality of threshold values set in stages for the finger image data A dynamic binarized image generation unit for obtaining a dark color area of
And a finger direction determining unit that determines a direction vector obtained by vector synthesis of the length directions of the dark color areas of each binarized image of the fingers as the finger direction of the finger image. Finger direction identification system. The finger direction identifying system according to claim 1,
The dynamic binarized image generation unit
A binarized image processing unit that sequentially binarizes the finger image data into different dark color areas based on a plurality of threshold values set in stages, and each dark color area that has been binarized An adjacent region specifying unit that monitors whether or not a proper state that is a binarized state immediately before any one of the dark region is connected to another adjacent dark region,
The finger direction specifying system, wherein the adjacent region specifying unit has a proper state holding function for acquiring and holding the binarized finger image when the proper state is obtained. In the finger direction identification system according to claim 2,
The binarized image processing unit has a stepwise threshold setting function for setting the threshold stepwise from the darkest color depth to a preset color depth when setting the threshold value. Finger direction identification system characterized by. In the finger direction identification system according to claim 2,
A processing range setting unit for adding an image frame of a predetermined size to finger image data input externally is provided in the data input stage for the finger image of the dynamic binarized image generation unit. Finger direction identification system characterized by. In the finger direction identification system according to claim 2,
A finger direction specifying system, wherein a smoothing processing unit that smoothes externally input finger image data is provided at a data input stage related to the finger image of the dynamic binary image generation unit. In the finger direction identification system according to any one of claims 1 to 5,
The finger direction determination unit
An area direction specifying function in which the length direction of the line segment from the centroid pixel of each dark color area as a starting point to the dark color pixel having the longest distance from the centroid pixel is the area direction; A finger direction determination system comprising: a finger direction determination processing function for determining a direction vector formed by vector composition of region directions as a finger direction of the finger image. In the finger direction identification system according to any one of claims 1 to 5,
The finger direction determination unit
While having a region direction specifying function in which the length direction of the line segment from the centroid pixel of each of the dark color regions generated to the dark color pixel having the longest distance from the centroid pixel is the region direction,
The finger direction specifying system, wherein the region direction in each dark color region is determined within a range of 180 degrees left and right on the extension side of each finger image. In the finger direction identification system according to any one of claims 1 to 5,
The finger direction determination unit
While having a finger direction determination processing function for determining a direction vector formed by vector composition of the region directions of each line segment as a finger direction of the finger image,
The region direction of the remaining line segments is vector-synthesized by excluding the line segment in the region direction where the difference between the finger direction specified by the finger direction determination processing function and the region direction of the original line segment is the largest. A finger direction specifying system comprising a composite direction updating function for updating a finger direction. In a finger direction identification system including a finger direction calculation processing unit that identifies a finger direction indicating a direction of a finger included in the finger image based on input finger image data by image processing,
Smooth the finger image data input externally,
Each finger portion of the finger image is binarized into a dark color state based on one specific threshold value among a plurality of threshold values set in a stepwise manner on the smoothed finger image data. To create a dark area in the proper state
The finger direction is characterized by vector-combining the length direction of the dark area of each binarized image of the fingers and determining the direction vector obtained by the vector composition as the finger direction of the finger image. Identification method. In the finger direction identification method according to claim 9,
In the step of binarizing each finger portion of the finger image into a dark color area,
In the binarization process of the finger image data, the entire finger image is sequentially binarized and displayed in different dark areas based on a plurality of threshold values set stepwise in advance,
The state of the binarized display is monitored so that an appropriate state immediately before any one of the dark color areas to be binarized displayed is connected to another adjacent dark color area is obtained. And
A finger direction specifying method characterized by specifying and holding a binarized finger image of the appropriate state when the appropriate state is obtained. The finger direction specifying method according to claim 9 or 10,
In the step of determining the finger direction of the finger image,
The area direction is the length direction of the line segment from the centroid pixel of each generated dark color area to the dark color pixel having the longest distance from the centroid pixel,
Thereafter, the finger direction specifying method is characterized in that a direction vector formed by vector synthesis of the region directions of each line segment is determined as the finger direction of the finger image. The finger direction specifying method according to claim 9 or 10,
In the finger direction determination step,
After determining the direction vector formed by vector composition of the area direction of each line segment as the finger direction of the finger image,
Excludes the line segment in the area direction where the difference between the determined finger direction and the area direction of each original line segment is the largest, and vector-combines the remaining line segment area directions to obtain the normal finger direction. A finger direction specifying method characterized by that. In a finger direction specifying system including a finger direction calculation processing unit that specifies a finger direction indicating a finger direction included in the finger image data based on input finger image data by image processing,
A smoothing function that smoothes finger image data that is input externally,
The smoothed finger image data is binarized into a dark state for each finger portion of the finger image based on one specific threshold value among a plurality of threshold values set in stages. Binarization processing function to generate dark areas in the proper state
And a finger direction determination processing function for determining a vector of the length direction of the dark color region of each binarized finger image and determining the direction vector obtained by the vector combination as the finger direction of the finger image. Provided,
A finger direction specifying program characterized in that each of these processing functions is realized by a computer provided in the finger direction calculation processing unit. In the finger direction identification program according to claim 13,
The binarization processing function
In the binarization processing of the finger image data, a binarization sequential processing function that sequentially binarizes the entire finger image into different dark areas for each of a plurality of threshold values set stepwise in advance,
Each of the fingers to be binarized so that an appropriate state immediately before any one of the dark color areas to be binarized is connected to another adjacent dark color area is obtained. Processing status monitoring function that monitors the status of the entire image,
And a binarized image specifying and holding function for specifying and holding a binarized finger image in the proper state when the appropriate state of each dark color region is obtained,
A finger direction specifying program characterized in that each of these functions is realized by the computer. In the finger direction identification program according to claim 13 or 14,
The finger direction determination processing function
An area direction specifying processing function in which the length direction of the line segment from the centroid pixel of each of the dark color areas as a starting point to the dark color pixel having the longest distance from the centroid pixel is the area direction;
A direction vector determination processing function for determining a direction vector formed by vector composition of the specified region directions of each line segment as a finger direction of the finger image;
A finger direction specifying program characterized in that each of these functions is realized by the computer. In the finger direction identification program according to claim 13 or 14,
The finger direction determination processing function
After determining the direction vector formed by vector composition of the area directions of each line segment as the finger direction of the finger image, the difference between the determined finger direction and the area direction of each original line segment is calculated, and the difference Specific line segment exclusion processing function that excludes the line segment that extends in the direction of the largest area,
And a finger direction specifying processing function for specifying a new finger direction obtained by vector synthesis of the region directions of the remaining line segments as a true finger direction instead of the determined finger direction,
And a finger direction specifying program characterized in that each of these functions is realized by the computer.

Images