JPH11250221A - Method and device for photographing facial picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and device for photographing a facial picture for always segmenting a facial picture from a proper position without being affected by any unnecessary picture due to a subject or noise components or the like. SOLUTION: This method includes a step for inputting the facial picture of a subject, step for operating the first segmenting position detection processing under a prescribed condition for detecting a segmenting position and a judgement reference at the time of detecting the segmenting position of the face of a subject from the inputted facial picture, step for operating the segmenting position re-detection processing by changing at least one of the prescribed condition for detecting the segmenting position and the judgement reference when the first segmenting position detection processing fails, and step for segmenting the face of the subject from the inputted facial picture.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a card issuing device for issuing a personal authentication card such as a cash card, a credit card, an employee ID card, or the like, and a face image to be printed on the personal authentication card. The present invention relates to a face image photographing method and a face image photographing device for photographing.

[0002]

2. Description of the Related Art For example, when issuing a personal identification card such as an employee ID card for identifying a person using a card issuing device that prints a face image or character information on a cash card size card, a CCD camera is used. Is printed on the card.

When photographing a face image of a cardholder, for example, the face image is cut out from an appropriate position from a photographing range of a CCD camera. At this time, the detection of the face position is performed by image processing. For example, by searching for the presence or absence of the head on a predetermined line from the input face image, the top of the face image and the right and left Is searched for, and the face image is cut out in a state where the face image always exists at a predetermined position within a predetermined frame whose length and width are specified.

In this case, a search method and a criterion for determining whether or not a head exists on a predetermined line are as described below. 1. Searching Method on Line The interval between pixels to be searched on the line is defined.

For example, if the interval is 12 with 500 pixels per line, the number of pixels searched on one line is about 41 pixels. 2. Judgment standard By judging a pixel different from the pixel of the background color as a pixel of the face (head), if a pixel considered to be the head on one line is equal to or more than the reference pixel number, the line is judged as a line having the head. I do.

For example, if the number of reference pixels is 6, and if the number of face (head) pixels on one line is 5 or less, the line is determined to be a line having no face (head), and
If the number of face (head) pixels on one line is 6 or more, the line is determined to be a line on which the face (head) exists.

[0007]

As described above, when photographing a face image of a cardholder, for example, a face image is cut out from an appropriate position from a photographing range of a CCD camera. The detection of the face position is performed by image processing. For example, in order to search for the presence or absence of a head on a predetermined line from an input face image, an interval (interval) of pixels to be searched is provided. If the number of pixels considered to be the head on one line is equal to or greater than the reference pixel number as a determination criterion, the line is determined to be the line where the head is located, and the top of the face image, the left and right temporal regions are determined. Is searched, and the image is cut out in a state where the face image always exists at a predetermined position in a predetermined frame whose length and width are specified.

However, in such a conventional face image photographing method, a face image is cut out from a photographing range of a CCD camera from an appropriate position, particularly when the photographed person has a gray hair or a headless hair. Often, they fail.

That is, in the face image photographing method according to the conventional technique as described above, a certain pixel (search pixel) of a certain line (search line) in the input face image is either a background or a face (head). Is determined based on the value of a color signal with a background color, but the subject has gray hair,
Since the color signal of the headless hair is close to the color signal of the background color, it has often failed to cut out the face image from an appropriate position from the shooting range of the CCD camera.

The failure to cut out such a face image from an appropriate position is caused, for example, when a noise component is present in the input face image and the noise component exists as an unnecessary image near the top of the head. In addition, the same can occur when the erroneous judgment is made on the top of the head.

Therefore, the present invention has been made in view of the above points, and when the color signal of the subject's head is close to the color signal of the background color, for example, if the subject has gray hair or headless hair Even if the head of the face image, or when there is an unnecessary image due to the noise component near the top of the face image,
The face image cannot be cut out from the proper position from the shooting range of the CCD camera without fail. It is an object of the present invention to provide a face image photographing method and a face image photographing device which can be cut out from the image.

[0012]

According to the present invention, in order to solve the above-mentioned problems, a step of inputting a face image of a subject and detecting a cut-out position of the face of the subject from the input face image. Performing a first cutout position detection process under predetermined conditions and criteria for detecting a cutout position; and, if the first cutout position detection process fails, the cutout position is determined. Changing at least one of the predetermined condition and the criterion to be detected and performing a re-detection process of the cutout position;
A step of extracting a face of the subject from the input face image based on the detected extraction position;
And a face image photographing method characterized by comprising:

According to the present invention, in order to solve the above-mentioned problems, a step of inputting a face image of a subject, and detecting a cutout position of the face of the subject from the input face image, Performing a first cutout position detection process under a predetermined condition and determination criterion for detecting a cutout position; and a predetermined process for detecting the cutout position when the first cutout position detection process fails. A step of sequentially changing at least one of the condition and the judgment criterion and performing a re-detection process of the cut-out position at a predetermined number of times until the detection process of the cut-out position succeeds; based on the detected cut-out position, Cutting out the face of the subject from the input face image.

According to the present invention, in order to solve the above-mentioned problems, a step of inputting a face image of a subject and a step of detecting a cut-out position of the face of the subject from the input face image, A step of performing a first detection process of the cutout position based on a predetermined search pixel interval number for detecting the cutout position and a predetermined reference pixel number; and a case where the first detection process of the cutout position fails. A predetermined number of search pixel intervals and a predetermined reference pixel number for detecting the cutout position, at least one of the determination criteria is sequentially changed, and the cutout position is detected at a predetermined number of times until the cutout position detection process is successful. Performing a re-detection process, and extracting a subject's face from the input face image based on the detected extraction position. Face image photographing method characterized by comprising the step, is provided.

According to the present invention, in order to solve the above-described problem, a step of inputting a face image of a subject, and detecting a cutout position of the face of the subject from the input face image, A step of performing a first detection process of the cutout position based on a predetermined search pixel interval number for detecting the cutout position and a predetermined reference pixel number; and a case where the first detection process of the cutout position fails. , The predetermined number of search pixel intervals for detecting the cutout position is halved, and the predetermined number of reference pixels is 2
Performing a re-detection process of the cut-out position a predetermined number of times until the detection process of the cut-out position is succeeded by successively changing the size to twice, based on the detected cut-out position,
Cutting out the face of the subject from the input face image.

According to the present invention, in order to solve the above-described problems, a step of inputting a face image of a subject and a step of detecting a cut-out position of the face of the subject from the input face image include: A step of performing a first detection process of the cutout position based on a predetermined search pixel interval number for detecting the cutout position and a predetermined reference pixel number; and a case where the first detection process of the cutout position fails. , The predetermined number of search pixel intervals for detecting the cutout position is halved, and the predetermined number of reference pixels is 2
Performing a re-detection process of the cut-out position up to twice by sequentially changing the cut-out position twice, and performing an error process when the re-detection process of the cut-out position up to two times fails;
When the re-detection processing of the cut-out position up to two times is successful, the step of cutting out the face of the subject from the input face image based on the detected cut-out position. The present invention provides a method of capturing a face image.

According to the present invention, in order to solve the above-mentioned problems, an input means for inputting a face image of a subject, and a cut-out position of the face of the subject from the face image input by the input means. Detection processing means for detecting the first cutout position under predetermined conditions and criteria for detecting the cutout position, and the first processing for detecting the cutout position by the detection processing means has failed. In this case, at least one of the predetermined condition and the criterion for detecting the cut-out position is changed to perform re-detection processing of the cut-out position, and the re-detection processing means detects the cut-out position. A face image capturing apparatus, comprising: a cutout means for cutting out a face of a subject from a face image input by the input means based on a cutout position. It is provided.

Further, according to the present invention, in order to solve the above-mentioned problems, an input means for inputting a face image of a subject, and a cut-out position of the face of the subject from the face image input by the input means. Detection processing means for detecting the first cutout position under predetermined conditions and criteria for detecting the cutout position, and the first processing for detecting the cutout position by the detection processing means has failed. In this case, at least one of the predetermined condition and the criterion for detecting the cut-out position is sequentially changed, and the re-detection process of the cut-out position is performed a predetermined number of times until the detection process of the cut-out position succeeds. Based on the detection processing means and the cut-out position detected by the re-detection processing means, the face of the subject is cut from the input face image input by the input means. Face image photographing apparatus characterized by comprising a clipping means for performing out, is provided.

Further, according to the present invention, in order to solve the above-described problems, an input means for inputting a face image of a subject, and a cut-out position of the face of the subject from the face image input by the input means. Detecting the first cutout position based on a predetermined search pixel interval number for detecting the cutout position and a predetermined reference pixel number, and a detection processing means for detecting the first cutout position. In the case where the detection processing of the second cutout position has failed, at least one of the conditions based on the predetermined search pixel interval number and the predetermined reference pixel number for detecting the cutout position is sequentially changed to detect the cutout position. Re-detection processing means for performing re-detection processing of the cut-out position a predetermined number of times until succeeding, and the cut-out position detected by the re-detection processing means Based on the input unit face image photographing apparatus characterized by comprising a clipping means, a to cut out of the face of the subject from O connexion inputted input face image is provided.

Further, according to the present invention, in order to solve the above-mentioned problems, an input means for inputting a face image of a subject, and a cut-out position of the face of the subject from the face image input by the input means. Detecting the first cutout position based on a predetermined search pixel interval number for detecting the cutout position and a predetermined reference pixel number, and a detection processing means for detecting the first cutout position. When the detection processing of the second cutout position fails, the predetermined search pixel interval number for detecting the cutout position is changed to １ ／ and the predetermined reference pixel number is sequentially doubled to detect the cutout position. Re-detection processing means for performing re-detection processing of the cutout position a predetermined number of times until succeeds, based on the cut-out position detected by this re-detection processing means, Face image photographing apparatus characterized by comprising a clipping means, the from O connexion inputted input face image to fill power means cuts out the face of the subject is provided.

Further, according to the present invention, in order to solve the above-mentioned problems, an input means for inputting a face image of a subject, and a cut-out position of the face of the subject from the face image input by the input means. Detecting the first cutout position based on a predetermined search pixel interval number for detecting the cutout position and a predetermined reference pixel number, and a detection processing means for detecting the first cutout position. If the detection processing of the second cutout position fails, the predetermined search pixel interval number for detecting the cutout position is reduced to 、, and the predetermined reference pixel number is sequentially changed to twice, and the cutout position is determined up to twice. Error processing means for performing error processing when the re-detection processing of up to two times by the re-detection processing means has failed, When the re-detection processing of the cut-out position up to two times by the step is successful, the cut-out for cutting out the face of the subject from the input face image input by the input means based on the detected cut-out position And a means for providing a face image photographing device.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 schematically shows a configuration of a face image photographing apparatus for photographing a face image to be printed on a personal authentication card in a card issuing apparatus for issuing a personal authentication card applied as an embodiment of the present invention. It is shown.

In FIG. 1, a CCD camera 1 as an image pickup means for picking up a face image of a subject is connected to an analog signal processing section 2, and a CRT display 3 as a display means is connected to the analog section 2. Have been.

This analog / digital signal processing unit 2
An image signal output as an analog signal from the CCD camera 1 is converted into a digital signal and sent to a first frame memory 4 described later, or image information of a digital signal sent from a second frame memory 5 described later is converted into an analog signal. It has a function of converting it into a signal and sending it to the display 3.

The analog / digital signal processing unit 2 is connected to a first frame memory 4 as a first storage unit and a second frame memory 5 as a second storage unit.

The first frame memory 4 has a CCD
The face image data captured by the camera 1 is stored, and the second
The frame memory 5 stores final face image data to be printed.

An address controller 6 is connected to the first frame memory 4 and the second frame memory 5. The first frame memory 4 and the second
The CPU (Central Processing Unit) 7 as control means is connected to the frame memory 5 and the address control unit 6.

The CPU 7 and the address control unit 6 control writing and reading of information to and from the first frame memory 4 and the second frame memory 5, and also control the first frame memory 4 to the second frame memory. 5 is controlled.

An operation keyboard 8 and a printing device 9 are connected to the CPU 7. The printing device 9 is connected to the second frame memory 5 and prints a face image sent from the second frame memory 5 at a predetermined position on the personal authentication card.

Next, the photographing process of the face image in such a configuration will be described with reference to the flowcharts shown in FIGS. FIG. 2 and FIG. 3 are flowcharts for explaining the face image capturing processing.

The subject's face image picked up by the CCD camera 1 is stored in the first frame memory 4 after A / D conversion (analog / digital conversion) by the analog / digital signal processing section 2.

At this time, the image picked up by the CCD camera 1 is displayed as a moving image on the display 3 in order to confirm the shooting state of the subject. At this time, the image data is stored in the first frame memory 4 in such a range that a sufficient space is formed in the peripheral portion even if the face is photographed large.

A second frame memory 5 provided separately from the first frame memory 4 for storing a captured face image
Stores a face image as a photographic image at the time of printing, and the transfer from the first frame memory 4 to the second frame memory 5 is performed by the CPU 7 which controls a photographing process.

Normally, the same image as the face image stored in the first frame memory 4 is displayed on the display 3, but after photographing, the face image of the subject to be printed is checked for confirmation. The face image in the second frame memory 5 is displayed.

The second frame memory 5 has an image having a size corresponding to the size and resolution of the face photograph to be printed. Now, when the photographing key for photographing the face image of the subject is pressed on the operation keyboard 8 (step S1), the photographing is started, and the moving image that has been written in the first frame memory 4 is write-protected. (Step S2), and conversely, reading of the frame memory is permitted.

At this time, for example, the display 3
The face image of the subject is not displayed. afterwards,
By reading the face image in the first frame memory 4, the CPU 7 detects the cutout position of the face image of the subject (Step S3a) and detects the size of the face of the subject (Step S3b).

That is, at this time, the first frame memory 4 is regarded as a two-dimensional array, and the CPU 7 uses the address control unit 6 to assign designated pixels in the H (horizontal) direction and the V (vertical) direction. Is managed so that it can be accessed.

The face image information read from the first frame memory 4 has a background imaged by the CCD camera 1 under a blue screen, for example, so that the cutout position of the face image of the subject can be easily detected. It is assumed that

Here, the details of the process of detecting the cut-out position of the face image of the subject (step S3a) will be described later. Then, a face size detection process performed after the detection of the cutout position of the face image of the subject (step S
3b) is performed as follows.

First, the CPU 7 scans the first frame memory 4 in the H direction from the top in the V direction to the bottom in the V direction, and detects a component having a color different from the background blue.
The address in the direction is the top position of the subject.

Subsequently, the CPU 7 scans the first frame memory 4 in the H direction, and detects a V component in which a color component different from the blue color of the background, which is equal to or larger than a pixel having a width in the H direction, is continuously detected. The address in the direction is the shoulder position of the subject.

The difference between the addresses of the head position and the shoulder position in the V direction, that is, as shown in FIG. 4, the pixel width from the head position to the shoulder position is used as the size of the face of the subject.

When the face size is detected in this way, the CPU
7 calculates the transfer magnification based on the detected face size (step S4). That is, the CPU 7 calculates the magnification at which the size of the face becomes the reference size, with the pixel at the top of the top position of the n pixels in the V direction in the first frame memory 4 as the upper end of the print image.

Next, the CPU 7 reduces and enlarges the face image in the first frame memory 4 based on the calculated magnification as shown in FIG.
And write (step S5).

When the transfer and write processing are completed, the CP
U7 displays the face image in the second frame memory 5 on the display 3 as a still image (step S6). The operator confirms in the face image displayed on the display 3 whether the face size, the cutout position is appropriate, and whether there is a problem such as blinking (Steps S7 and S7).
8, S9) If the size of the face and the cut-out position are appropriate and there are no problems such as blinking, the enter key in the operation keyboard 8 is pressed (step S10).

If the result of this check is that there is a blink, the flow returns to step S1 and the photographing key is pressed again to perform photographing again. If there is an erroneous detection of the size of the face or the cutout position due to the hairstyle or clothing of the subject, it is corrected by the correction means described below.

For example, when the size of the face displayed on the display 3 is too large, the CPU 7 presses, for example, a minus (-) key in the operation keyboard 8 (step S11), and the CPU 7 sets the first key. The magnification is recalculated so that the transfer magnification from the frame memory 4 to the second frame memory 5 becomes smaller by a certain rate than the previous time (step S12).

Thereafter, the flow returns to step S5, and the transfer and the writing are controlled again based on the recalculated magnification, thereby correcting the face size to be reduced. During the transfer and writing control, the display of the face image on the display 3 is prohibited, and after the transfer and writing are completed, the face image in the second frame memory 5 is displayed again.

On the other hand, if the size of the face is too small, the CPU 7 presses, for example, a plus (+) key in the operation keyboard 8 (step S11), and the CPU 7
The magnification is re-calculated so that the transfer magnification from the first frame memory 4 to the second frame memory 5 becomes larger by a certain rate than the previous time (step S12).

Thereafter, the flow returns to step S5, and the transfer and writing control are performed again based on the recalculated magnification, thereby correcting the face size to be large. If the cut-out position of the face image is inappropriate, the correction is performed by, for example, up and down arrow keys in the operation keyboard 8.

That is, when the face position is slightly up,
By pressing the down (↓) arrow key in the operation keyboard 8, the CPU 7 corrects the data by performing the transfer and writing again so that the face position is in the downward direction at the previously set magnification. (Steps S13 to S1)
6).

Similarly, when the face position is slightly lower, the correction is made by pressing the upward (↑) arrow key in the operation keyboard 8. After these corrections, the enter key in the operation keyboard 8 is pressed (step S10).

As a result, the CPU 7 transfers the face image in the second frame memory 5 to the printing device 9. The printing device 9 prints the face image and predetermined character information transferred from the second frame memory 5 on a card.

Next, in the above-described face image photographing process, the details of the process of detecting the cut-out position of the face image of the subject (step S3a) as a main part of the present invention will be described with reference to the search principle diagram shown in FIG. This will be described with reference to the flowchart shown in FIG.

FIG. 5 is a diagram for explaining the principle of a search method for detecting a cut-out position of a face image employed in a process of detecting a cut-out position of a face image of a subject as a main part of the present invention.

FIG. 6 is a flowchart for explaining a process of detecting a cut-out position of a face image of a subject, which is an essential part of the present invention, in the process of photographing a face image. As described above, the detection of the cutout position of the face image of the subject is performed by:
This is performed under a predetermined search method and judgment criteria. For example, a search is performed for whether or not a head exists on a predetermined line from an input face image, and thereby the top and right and left sides of the face image are searched. The head is searched and cut out in a state where the face image always exists at a predetermined position within a predetermined frame whose length and width are specified.

First, the principle of the search method for detecting the cut-out position of the face image employed in the detection processing of the cut-out position of the face image of the subject, which is a main part of the present invention, will be described. Although only the H-line direction at the top of the head will be described here, the same applies to the V-line direction.

That is, the principle of this search method is that, as shown in FIG. 5, a range in which a search line is searched is sequentially divided into two by a so-called binary search or a search method called a binary search method (binary search). By setting lines (1, 2, 3,..., N) in the drawing, a line where the target vertex exists is found.

That is, first, a pixel different from the background color pixel is determined as a face pixel by being set to the center line (shown in FIG. 1) that bisects the H line direction, so that a line on the one line is determined. Since the number of pixels considered to be a face is equal to or larger than the reference pixel number, it is determined that the target vertex exists in a line above that line.

Then, a center line (shown in FIG. 1) that bisects the H line direction and a line (shown in FIG. 2) that bisects the top end in the H line direction are set. Since there is no pixel different from the background color pixel, it is determined that the target vertex exists in a line below that line.

Then, next, a line (shown in FIG. 3) which bisects the line shown in FIG. 2 and the line shown in FIG. 1 is set. Therefore, it is determined that the target vertex exists in a line above that line.

Then, a line (shown in FIG. 4) that bisects the line shown in FIG. 3 and the line shown in FIG. 2 is set. Since there is no line, it is determined that the target vertex exists in a line below that line.

In the following, if the number of pixels considered to be the head on the line set in the same manner is equal to or more than the reference number of pixels, it is determined that the target vertex exists on the line above the line, and the setting is performed. If the number of pixels considered to be the head on the set line is not equal to or greater than the reference number of pixels, it is determined that the target vertex exists on the line below the line, and the final sequence is repeated. ,
It is possible to find the line in the H-line direction where the target vertex exists.

Then, the H in which the target crown exists
Based on the line information in the line direction and the line information in the V-line direction in which the left and right temporal regions found in the same manner exist, the vertical and horizontal lengths indicated by broken lines in FIG. 4 are designated. The predetermined frame is set as a cutout position, and the face image can be cut out in a state where the face image always exists at a predetermined position in the frame.

Next, a description will be given of a search method and a criterion for searching for whether or not a head exists on a predetermined line actually employed in the present invention, based on the above principle. As a search method employed in the present invention, an interval between pixels to be searched on a line is defined, and a pixel different from a background color pixel is determined as a face (head) pixel as a criterion. If, on one line, the number of pixels considered to be the head is equal to or more than the reference number of pixels, the line itself is determined to be the line where the head exists, similar to the conventional one.

However, in addition to the above-described search method and determination criterion, a predetermined number of times (for example, 2
Re-detection processing is performed up to a total of three times, and the number of intervals in the re-detection processing is sequentially smaller than the number of intervals of the first search (for example, 1 /
2) By increasing the number of reference pixels sequentially (for example, twice) from the number of reference pixels in the first search, it is possible to prevent a failure in cutting out a face image from an appropriate position, The search can be performed more reliably.

That is, as shown in FIG.
In step S21, a face detection process by a first search is performed based on a predetermined search method and a determination criterion as described later.
At 22, it is determined whether or not the face detection process by the first search is successful.

If the face detection process by the first search is successful, the process is terminated as it is. If the face detection process by the first search is unsuccessful, the process proceeds to the re-detection process after step S23.

In step S23, it is determined whether or not the face detection processing has been performed a predetermined number of times or more. If the face detection processing has been performed a predetermined number of times, it is treated as an error processing. If the determination in step S23 is that the face detection processing is not equal to or more than the predetermined number of times, in step S24, the number of intervals for the first (or second) search is smaller (for example, 1).
/ 2) and the number of reference pixels is set to the first (or second)
The search method and the criterion are changed so that the number of search pixels is larger (for example, twice) than the reference pixel number of the search.

Then, the face detection processing by the second and subsequent searches is performed in step S21 based on the changed search method and determination criteria, and the same determination and processing as described above are performed.

Next, an example under the following conditions will be described. Re-detection processing: 2 times First search method and criteria 1 line: 27 pixels Interval interval: 4 Reference pixel number: 3 When the value of the background pixel is 100, 130 or more and -30 or less and 70 or less Is the face (head) pixel.

As shown in FIG. 7, in the first search,
If the square pixel shown is a search pixel, there is only one pixel (pixel value 68) considered to be a face in the seven search pixels, and the face detection fails because the reference pixel number 3 is not satisfied.

Redetection Processing (First Time) The interval is set to 2 and the reference surface prime number is set to 6. Assuming that the square pixel shown is a search pixel, five pixels (pixel values 68, 60, 50,
68, 65), which does not satisfy the reference pixel number 6, so that face detection fails.

Redetection processing (second time) The interval is set to 1 and the reference pixel number is set to 12. Assuming that the square pixel shown is a search pixel, 13 pixels (a pixel value of 68, 60, 5) considered to be a face are included in 27 search pixels.
0,60,55,60,50,67,68,69,6
8, 65, 64), which satisfies the reference pixel number 12, so that face detection is successful.

In this way, every time the re-detection process is performed, the number of intervals is reduced to half, the number of search plane elements is increased, and the number of reference pixels is doubled to make the criterion strict. By reducing the number of intervals to が, the number of search pixels increases and the processing time increases, but the search can be performed more reliably.

Although the number of reference pixels is doubled, a line where the target vertex exists finally is found by the above-described search method called binary search or binary search method (binary search). When going, if there is a face (head) on the final search line, there should be no problem even if the criterion is strict.

That is, in the prior art, the interval between pixels to be searched on a line is defined as a search method, and a pixel different from a background color pixel is determined as a face (head) pixel as a criterion. By judgment, 1
If the number of pixels that are considered to be the head on the line is equal to or greater than the reference number of pixels, it is only determined that the line is the line where the head exists. If not, the face detection fails directly. On the other hand, in the present invention, each time the re-detection process is performed, the number of intervals is halved to increase the number of search plane primes and the number of reference pixels is increased. By changing the search method and the criterion so as to make the criterion stricter by doubling, the number of search pixels increases and the processing time becomes longer by reducing the number of intervals to 、, but the search is more reliably performed. Therefore, it is possible to eliminate a failure to cut out a face image from an appropriate position, and to search more reliably.

[0078]

As described above in detail, according to the present invention, when the color signal of the subject's head is close to the color signal of the background color, for example, the subject has gray or headless hair. In some cases, or when there is an unnecessary image due to a noise component near the top of the face image, it fails to cut out the face image from an appropriate position from the shooting range of the CCD camera. Provided is a face image photographing method and a face image photographing apparatus capable of always cutting out a face image from an appropriate position without being affected by an unnecessary image due to a subject, a noise component, or the like. be able to.

[Brief description of the drawings]

FIG. 1 is a configuration of a face image photographing device for photographing a face image to be printed on a personal authentication card in a card issuing device that issues a personal authentication card applied as an embodiment of the present invention; Is schematically shown.

FIG. 2 is a flowchart illustrating a face image photographing process performed by the face image photographing apparatus of FIG. 1;

FIG. 3 is a flowchart for explaining a face image photographing process by the face image photographing apparatus of FIG. 1;

FIG. 4 is a diagram for explaining a face image cutout position detection, reduction, and enlargement process performed by the face image photographing apparatus of FIG. 1;

FIG. 5 is a diagram for explaining the principle of a search method for detecting a cut-out position of a face image employed in a detection process of a cut-out position of a face image of a subject as a main part of the present invention.

FIG. 6 is a flowchart for explaining a process of detecting a cut-out position of a face image of a subject, which is a main part of the present invention, in a face image shooting process;

FIG. 7 is a diagram for explaining a process of detecting a cut-out position of a face image of a subject as a main part of the present invention in a process of capturing a face image;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Camera (imaging means), 2 ... Analog / digital signal processing part, 3 ... Display (display means), 4 ... 1st frame memory (1st storage means), 5 ... 2nd frame memory (2nd 6) Address control section, 7 CPU (control means), 8 Operation keyboard, 9 Printing device.

Claims (10)

[Claims] 1. A step of inputting a face image of a subject, and detecting a cutout position of the face of the subject from the input face image for the first time under a predetermined condition and a criterion for detecting the cutout position. Detecting the cut-out position, and, when the first cut-out position detection process fails, changing at least one of the predetermined conditions and criteria for detecting the cut-out position and cutting out the cut-out position. A face image capturing method, comprising: performing a position re-detection process; and performing a clipping of a face of a subject from the input face image based on the detected clipping position. 2. A step of inputting a face image of a subject, and detecting a cutout position of the face of the subject from the input face image for a first time under a predetermined condition and a criterion for detecting the cutout position. Performing the detection process of the cut-out position, and, if the first detection process of the cut-out position fails, sequentially changing at least one of the predetermined conditions and criteria for detecting the cut-out position. Performing a re-detection process of the cut-out position a predetermined number of times until the detection process of the cut-out position succeeds; and, based on the detected cut-out position, performing a cut-out of the face of the subject from the input face image. A face image capturing method, comprising: 3. A step of inputting a face image of a subject, and when detecting a cutout position of the face of the subject from the input face image, a predetermined number of search pixel intervals for detecting the cutout position and a predetermined reference pixel Performing a first cutout position detection process based on a criterion based on a number; and, if the first cutout position detection process fails, a predetermined number of search pixel intervals for detecting the cutout position and a predetermined A step of sequentially changing at least one of the criteria based on the number of reference pixels and performing a re-detection process of the cut-out position a predetermined number of times until the detection process of the cut-out position succeeds; and Extracting a face of the subject from the input face image. 4. A step of inputting a face image of a subject, and when detecting a cut-out position of the face of the subject from the input face image, a predetermined number of search pixel intervals for detecting the cut-out position and a predetermined reference pixel Performing a first cutout position detection process based on a criterion based on the number; and, if the first cutout position detection process fails, the predetermined search pixel interval number for detecting the cutout position is set to 1 / (2) performing a re-detection process of the cut-out position a predetermined number of times until the detection process of the cut-out position is succeeded by sequentially changing the predetermined number of reference pixels to twice, and based on the detected cut-out position. Extracting a face of a subject from the input face image. 5. A step of inputting a face image of a subject, and when detecting a cut-out position of the face of the subject from the input face image, a predetermined number of search pixel intervals for detecting the cut-out position and a predetermined reference pixel Performing a first cutout position detection process based on a criterion based on the number; and, if the first cutout position detection process fails, the predetermined search pixel interval number for detecting the cutout position is set to 1 / 2 and sequentially changing the predetermined number of reference pixels to twice, and performing re-detection processing of the cut-out position up to twice, and performing error processing when the re-detection processing of the cut-out position up to two times fails. And, when the re-detection processing of the cut-out position up to two times is successful, the subject is extracted from the input face image based on the detected cut-out position. Face image photographing method characterized by comprising the step of cutting out a face, the. 6. An input means for inputting a face image of a subject, and a predetermined condition and judgment for detecting the cut-out position when detecting a cut-out position of the face of the subject from the face image input by the input means. Detection processing means for detecting a first cutout position based on a reference; predetermined conditions and criteria for detecting the cutout position when the first detection processing of the cutout position by the detection processing means fails; Re-detection processing means for re-detecting the cut-out position by changing at least one of them; and a face input by the input means based on the cut-out position detected by the re-detection processing means. A face image capturing apparatus, comprising: a clipping unit that clips a face of a subject from an image. 7. An input means for inputting a face image of a subject, and a predetermined condition and determination for detecting the cut-out position when detecting the cut-out position of the face of the subject from the face image input by the input means. Detection processing means for detecting a first cutout position based on a reference; predetermined conditions and criteria for detecting the cutout position when the first detection processing of the cutout position by the detection processing means fails; Re-detection processing means for re-detecting the cut-out position a predetermined number of times until at least one of them is sequentially changed and the detection processing of the cut-out position succeeds, and the cut-out position detected by the re-detection processing means And a cutout means for cutting out the face of the subject from the input face image input by the input means based on Image capturing device. 8. An input means for inputting a face image of a subject, and a predetermined search pixel interval for detecting the cut-out position when detecting the cut-out position of the face of the subject from the face image input by the input means. Detection processing means for performing detection processing of a first cutout position based on a criterion based on the number of pixels and a predetermined reference pixel number; and detecting the cutout position when the first detection processing by the detection processing means fails. The re-detection process of the cutout position is performed at a predetermined number of times until at least one of the conditions based on the predetermined search pixel interval number and the predetermined reference pixel number for detecting is detected and the cutout position detection process succeeds. Re-detection processing means; and an input face image input by the input means based on the cutout position detected by the re-detection processing means. A cutout means for cutting out a face of a subject from an image. 9. An input means for inputting a face image of a subject, and a predetermined search pixel interval for detecting the cut-out position when detecting the cut-out position of the face of the subject from the face image input by the input means. Detection processing means for performing detection processing of a first cutout position based on a criterion based on the number of pixels and a predetermined reference pixel number; and detecting the cutout position when the first detection processing by the detection processing means fails. The re-detection process of the cut-out position is performed a predetermined number of times until the predetermined process of detecting the cut-out position is succeeded by sequentially changing the predetermined number of search pixel intervals for detecting the pixel to 1/2 and the predetermined number of reference pixels twice. Re-detection processing means; and a subject based on the cut-out position detected by the re-detection processing means, based on the input face image input by the input means. And a cutout means for cutting out the face. 10. An input means for inputting a face image of a subject, and a predetermined search pixel interval for detecting the cut-out position when detecting a cut-out position of the face of the subject from the face image input by the input means. Detection processing means for performing detection processing of a first cutout position based on a criterion based on the number of pixels and a predetermined reference pixel number; and detecting the first cutout position by the detection processing means when the detection processing of the first cutout position fails. The predetermined number of search pixel intervals for detecting and the predetermined number of reference pixels are sequentially changed to twice, and re-detection processing of the cutout position is performed up to twice by the re-detection processing means. Error processing means for performing error processing when re-detection processing of the cut-out position has failed, and re-detection processing of the cut-out position by this re-detection processing means has been successful up to twice. A face image photographing apparatus, comprising: a cutout means for cutting out a face of a subject from a face image input by the input means based on the detected cutout position. .