JP2006166348A - Face image photographing apparatus, face image decision aparatus and face image deciding program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable the decision of the quality of a face image from image data of one frame for every image data and also the decision from mutual discontinuous image data taken at random timings. <P>SOLUTION: The face image photographing apparatus has: an imaging unit 120 for generating and outputting image data by using an imaging device driven in the interlace system for imaging the face; a pattern detector 160 for detecting a specified pattern at a specified portion of the image data of the face image generated by the imaging unit; and a decision unit 180 for deciding the quality of the face image according to the detection result by the detector. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a face image capturing device, a face image determining device, and a face image determining program for capturing a face image, and more particularly to a face image capturing device and a face image capable of determining pass / fail of a captured face image. The present invention relates to a determination device and a face image determination program.

A face photograph may be used for various identification cards and the like. In this case, a face photograph taken by a face image photographing system is used.
In this type of face image capturing system, a video signal or image data generated by an image capturing unit having an image sensor that is driven by a 2: 1 interlace used in a television system or the like is used. There is.

  By the way, in the face image photographing system, it is important that the photographed face image is not an instant such as a blink. For this reason, it is necessary to determine whether the face image captured by the face image capturing system is blinking.

For example, the following Patent Document 1 describes a technique for detecting whether there is a motion in an image or whether the image is stationary. Patent Document 2 below describes a method for detecting blinking of a face image.
JP 63-42294 (first page, FIG. 1) JP 2001-67459 A (first page, FIG. 1)

  The technique described in Patent Literature 1 described above detects motion in an image using frame correlation. Therefore, a plurality of continuous images that are periodically repeated like a television video signal are required. For this reason, motion detection cannot be applied to discontinuous image data such as when captured by a camera.

  The technique described in Patent Document 2 also detects movement in an image such as blinking by comparing two consecutive frames of image data. For this reason, it is necessary to continuously repeat images of a plurality of frames like a television video signal. Furthermore, as described in the specification of Patent Document 2, “shooting can be completed when the face state desired by the photographer after the second frame is reached” is described. There was a problem that it was not possible to make judgment based on image data.

  As described above, in any patent document, continuous image data of a plurality of frames such as a television system video signal is necessary, and a comparison of a plurality of frames is necessary. As a result, at random timing, There has been a problem that discontinuous and nonconsecutive image data cannot be compared / determined as when taken with a camera.

The present invention has been made in view of the above problems in the prior art,
When determining the quality of a face image without requiring continuous image data of a plurality of frames, it is possible to determine with one frame of image data for each image data, and a discontinuous image taken at random timing It is an object of the present invention to provide a face image photographing device, a face image determining device, and a face image determining program capable of determining data.

That is, the above-described problems are solved by the inventions listed below.
(1) According to the first aspect of the present invention, an imaging unit that generates and outputs image data by imaging with an imaging device driven by an interlace method, and image data of a face image generated by the imaging unit A pattern detection unit that detects a specific pattern in a predetermined part of the image, and a determination unit that determines the quality of the face image based on the detection result of the pattern detection unit. (2) The invention according to claim 2 In the face image photographing device of (1), a part recognition unit for recognizing a predetermined part of the face image is provided, and the part recognition unit is arranged in a direction orthogonal to the driving direction of the interlace method of the image sensor. A dimensional filter is provided, and the predetermined portion is detected by a change in a signal value by the filter.

  (3) The invention according to claim 3 is the face image capturing device according to (1) or (2), wherein the pattern detection unit is a one-dimensional direction orthogonal to an interlaced drive direction of the image sensor. The specific pattern in the predetermined part is detected by the change of the signal value by the filter.

  (4) The invention according to claim 4 is the face image capturing device according to (3), wherein the filter is a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor. It is characterized by comprising.

  (5) In the face image capturing device according to any one of (1) to (4), the determination unit compares the detection result of the pattern detection unit with a predetermined threshold value. The image data whose signal value change is not more than a threshold value is determined to be good.

  (6) In the face image capturing device according to any one of (1) to (5), the output of the image data is determined according to whether the determination result by the determination unit is good. Selecting means, wherein the selecting means executes the detection on the plurality of stored image data by the pattern detection unit, and according to the detection result on the plurality of image data by the pattern detection unit. When the determination is executed by the determination unit, whether to output each image data is determined according to the determination result of the plurality of image data.

  (7) The invention according to claim 7 is directed to a case where the specific pattern is detected in any one of a storage unit storing a series of image data and the image data stored in the storage unit. A pattern distribution creation unit that creates a pixel position where a pattern is generated as a pattern distribution, and the pattern detection unit is a pixel indicated by the pattern distribution with respect to other image data in the series of image data. The face image determination apparatus according to claim 8, wherein pattern detection is performed at a position.

  (8) The invention according to claim 8 is a signal input unit that receives supply of image data generated by imaging a face with an image sensor that is driven in an interlaced manner, and is input to the signal input unit A face image determination apparatus comprising: a pattern detection unit that detects a specific pattern in a predetermined part of image data of a face image; and a determination unit that determines the quality of a face image based on a detection result of the pattern detection unit. is there.

  (9) The invention according to claim 9 is a signal input unit that receives supply of an interlaced video signal generated by capturing an image of a face with an image sensor driven by an interlaced method, and the signal input unit A frame data generation unit that generates image data of one frame from the odd field and the even field of the video signal input to the video signal; and a specific part of a face image of the face image of the image data generated by the frame data generation unit The face image determination apparatus includes a pattern detection unit that detects a pattern, and a determination unit that determines the quality of the face image based on the detection result of the pattern detection unit.

  (10) The invention according to claim 10 is the face image determination device according to (7) or (8), further comprising a part recognition unit for recognizing a predetermined part of the face image, wherein the part recognition unit includes the imaging element. The interlaced drive direction is orthogonal to the one-dimensional direction of the filter, and the predetermined portion is detected by a change in signal value using the filter.

  (11) The invention according to claim 11 is the face image determination device according to any one of (7) to (9), wherein the pattern detection unit is in a direction orthogonal to the driving direction of the interlace method of the image sensor. A one-dimensional filter is provided, and the predetermined portion is detected by a change in signal value by the filter.

  (12) The invention according to claim 12 is the face image determination device according to (10), wherein the filter is a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor. It is characterized by comprising.

  (13) The invention according to claim 13 is the face image determination device according to any one of (7) to (11), wherein the determination unit compares the detection result of the pattern detection unit with a predetermined threshold value. The image data whose signal value change is not more than a threshold value is determined to be good.

  (14) According to the fourteenth aspect of the present invention, in the face image determination device according to any one of (7) to (12), whether or not the image data can be output is determined according to the determination result of the determination unit. Selecting means, wherein the selecting means executes the detection on the plurality of stored image data by the pattern detection unit, and according to the detection result on the plurality of image data by the pattern detection unit. When the determination is executed by the determination unit, whether to output each image data is determined according to the determination result of the plurality of image data.

  (15) In the invention described in claim 15, when the specific pattern is detected in any one of a storage unit that stores a series of a plurality of image data and the image data stored in the storage unit, the specification is performed. A pattern distribution creation unit that creates a pixel position where a pattern is generated as a pattern distribution, and the pattern detection unit is a pixel indicated by the pattern distribution with respect to other image data in the series of image data. The face image determination apparatus according to claim 8, wherein pattern detection is performed at a position.

  (16) The invention according to claim 16 is an input routine for receiving supply of image data generated by imaging a face with an image sensor that is driven in an interlaced manner, and specifying a predetermined portion of the image data This is a face image determination program having a detection routine for detecting the pattern and a determination routine for determining pass / fail of the face image based on the detection result in the detection routine.

As described above, according to the present invention described in each claim, the following effects can be obtained.
(1) According to the invention of the face image photographing device of the first aspect, image data is generated by taking an image with an imaging device driven by an interlace method, and a specific part of the image data of the face image is specified at a predetermined part. A pattern is detected, and the quality of the face image is determined based on the detection result of the specific pattern.

  The predetermined part is an eye part of the face image. The specific pattern is a striped pattern appearing in a frame image composed of a field image with eyes open and a field image with eyes closed. Alternatively, the specific pattern is a striped pattern appearing in a frame image formed by field images having different eye opening degrees. Here, the stripe pattern occurs in the scanning line direction of the interlace method.

  That is, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and therefore the quality of the face image is determined without requiring continuous image data of a plurality of frames. In this case, each image data can be determined by one frame of image data, and not only a series of image data continuously captured but also discontinuous image data captured at random timings. It is also possible to determine.

  (2) In the invention according to claim 2, when the face image capturing device of (1) recognizes a predetermined part of the face image, it is one-dimensional in a direction orthogonal to the driving direction of the interlace method of the image sensor. A direction filter is provided, and a predetermined portion is detected by a change in signal value by the filter.

  Here, since a predetermined part is detected by a one-dimensional filter in a direction orthogonal to the interlaced driving direction of the image sensor, interlaced scanning is performed on the eye part of the face image. A specific pattern (striped pattern) generated in the line direction can be executed at high speed with a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required, and the quality of a face image is determined. In addition, each image data can be determined by one frame of image data, not only between a series of image data captured continuously but also between discontinuous image data captured at random timing. Judgment can also be executed reliably.

  (3) In the invention according to claim 3, when detecting a specific pattern in a predetermined part in the face image capturing device according to (1) or (2), it is orthogonal to the driving direction of the interlace method of the image sensor. And a specific pattern in a predetermined part is detected by a change in the signal value.

  Here, since a specific pattern is detected at a predetermined site by a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor, it occurs in the interlaced scanning line direction. A specific pattern (striped pattern) can be executed at high speed by a simple process with a small amount of computation, and each image is not necessary when determining the quality of a face image without requiring continuous image data of a plurality of frames. Judgment can be made with one frame of image data for each data, and not only judgment of a series of image data taken continuously but also judgment of discontinuous image data shot at random timing is ensured. It becomes possible to execute.

  Here, the detection of a predetermined part and the detection of a specific pattern in the predetermined part are performed by a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor. Therefore, it is possible to execute a specific pattern (striped pattern) generated in the interlace scanning line direction at a high speed by a simple process with a small amount of calculation while further reducing the circuit scale and the number of frames. When determining the quality of a face image without the need of the image data, it is possible to determine with one frame of image data for each image data, and only by determining a series of image data taken continuously. In addition, it is possible to reliably execute determination of discontinuous image data captured at random timing.

  (4) In the invention described in claim 4, in the face image photographing device of (3), the signal value is changed by a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor. To detect a predetermined part.

  Here, since a specific pattern is detected in a predetermined part by a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor, the interlace method scanning is performed. A specific pattern (striped pattern) generated in the line direction can be executed at high speed with a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required, and the quality of a face image is determined. In addition, each image data can be determined by one frame of image data, not only between a series of image data captured continuously but also between discontinuous image data captured at random timing. Judgment can also be executed reliably.

  In addition, here, detection of a predetermined part and detection of a specific pattern in the predetermined part are performed by a filter of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the interlace method of the image sensor. Therefore, while further reducing the circuit scale and the specific pattern (striped pattern) generated in the scanning line direction of the interlace method, it can be executed at high speed by a simple process with a small amount of calculation, When determining the quality of a face image without requiring continuous image data of a plurality of frames, determination can be made with one frame of image data for each image data. In addition to this determination, it is possible to reliably execute determination of discontinuous image data captured at random timing.

  (5) In the invention according to claim 5, in the face image photographing device according to any one of (1) to (4), a detection result of a specific pattern in a predetermined part of the image data of the face image is set to a predetermined threshold value. And the image data whose signal value change is not more than the threshold value is determined to be good.

  In other words, the quality of the face image is determined by comparing the signal value, which is the detection result of a specific pattern at a predetermined part in each image data, with a predetermined threshold value. When determining the quality of a face image without requiring data, it is possible to determine with one frame of image data for each image data, not only for determination of a series of image data taken continuously, It is also possible to determine discontinuous image data captured at random timing.

  (6) In the invention according to claim 6, in the face image photographing device according to any one of (1) to (5), detection of a specific pattern in a predetermined part is executed for a plurality of stored image data. The determination unit executes determination according to the detection results for the plurality of image data, and determines whether each image data can be output according to the determination result of the plurality of image data.

  In other words, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and whether or not each image data can be output is determined according to the quality of the determination results of the plurality of image data. Therefore, it is possible to determine the quality of a face image without using multiple frames of continuous image data, and it is possible to determine with one frame of image data for each image data. In addition to the determination of a series of image data, it is possible to determine discontinuous image data captured at random timing, and it is also possible to select which of a plurality of image data is to be output.

  (7) In the invention according to claim 7, in the face image photographing device according to any one of (1) to (6), a series of a plurality of image data is stored, and any one of the stored series of image data is stored. When the specific pattern is detected, the pixel position where the specific pattern is generated is created as a pattern distribution, and the other image data in the series of image data is indicated by the pattern distribution. Pattern detection is performed at the pixel position.

  That is, since the position where the blink occurs in the image data in which the blink is detected is obtained as a pattern distribution, and the pattern detection is performed on the remaining image data of the series of image data with respect to the pattern distribution, Of the series of images at the timing before and after the blink, those that do not detect the blink are determined as the high-quality images. Therefore, it is possible to reliably detect the high-quality images that do not include the blink. Further, the moment when the subject blinks is detected and identified, the region of both eyes is identified by the distribution of the specific pattern of the blinked image data, and other image data is detected for the identified region. Therefore, it is possible to more reliably detect a high-quality image that does not include blinking.

  (8) In the face image determination device according to the eighth aspect of the present invention, when image data generated by taking an image with an imaging device driven by an interlace method is supplied, the image data of the face image is received. A specific pattern in a predetermined part of the image is detected, and the quality of the face image is determined based on the detection result of the specific pattern.

  The predetermined part is an eye part of the face image. The specific pattern is a striped pattern appearing in a frame image composed of a field image with eyes open and a field image with eyes closed. Alternatively, the specific pattern is a striped pattern appearing in a frame image formed by field images having different eye opening degrees. Here, the stripe pattern occurs in the scanning line direction of the interlace method.

  That is, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and therefore the quality of the face image is determined without requiring continuous image data of a plurality of frames. In this case, each image data can be determined by one frame of image data, and not only a series of image data continuously captured but also discontinuous image data captured at random timings. It is also possible to determine.

  (9) In the invention of the face image determination device according to claim 9, when receiving an interlaced video signal generated by imaging a face with an imaging device driven by an interlace method, Image data of a face image of one frame is generated from an odd field and an even field of the input video signal, a specific pattern in a predetermined part of the image data of the face image is detected, and a detection result of the specific pattern The quality of the face image is determined based on the above.

  The predetermined part is an eye part of the face image. The specific pattern is a striped pattern appearing in a frame image composed of a field image with eyes open and a field image with eyes closed. Alternatively, the specific pattern is a striped pattern appearing in a frame image formed by field images having different eye opening degrees. Here, the stripe pattern occurs in the scanning line direction of the interlace method.

  That is, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and therefore the quality of the face image is determined without requiring continuous image data of a plurality of frames. In this case, each image data can be determined by one frame of image data, and not only a series of image data captured continuously but also a discontinuous video signal captured at random timing, It is also possible to determine between image data.

  (10) In the invention according to claim 10, in the face image determination device according to (7) or (8), when the predetermined part of the face image is recognized, the driving direction of the interlace method of the image sensor is orthogonal. A one-dimensional direction filter is provided, and a predetermined portion is detected by a signal value change by the filter.

  Here, since a predetermined part is detected by a one-dimensional filter in a direction orthogonal to the interlaced driving direction of the image sensor, interlaced scanning is performed on the eye part of the face image. A specific pattern (striped pattern) generated in the line direction can be executed at high speed with a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required, and the quality of a face image is determined. In addition, each image data can be determined by one frame of image data, not only between a series of image data captured continuously but also between discontinuous image data captured at random timing. Judgment can also be executed reliably.

  (11) In the invention according to claim 11, in the face image determination device according to any one of (7) to (9), when detecting a specific pattern in a predetermined part, the driving direction of the interlace method of the image sensor Is provided with a one-dimensional filter in a direction orthogonal thereto, and a predetermined part is detected by a change in signal value by the filter.

  Here, since a specific pattern is detected at a predetermined site by a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor, it occurs in the interlaced scanning line direction. A specific pattern (striped pattern) can be executed at high speed by a simple process with a small amount of computation, and each image is not necessary when determining the quality of a face image without requiring continuous image data of a plurality of frames. Judgment can be made with one frame of image data for each data, and not only judgment of a series of image data taken continuously but also judgment of discontinuous image data shot at random timing is ensured. It becomes possible to execute.

  Here, the detection of a predetermined part and the detection of a specific pattern in the predetermined part are performed by a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor. Therefore, it is possible to execute a specific pattern (striped pattern) generated in the interlace scanning line direction at a high speed by a simple process with a small amount of calculation while further reducing the circuit scale and the number of frames. When determining the quality of a face image without the need of the image data, it is possible to determine with one frame of image data for each image data, and only by determining a series of image data taken continuously. In addition, it is possible to reliably execute determination of discontinuous image data captured at random timing.

  (12) In the invention according to claim 12, in the face image determination device according to (10), the signal value is changed by a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor. To detect a predetermined part.

  Here, since a specific pattern is detected in a predetermined part by a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor, the interlace method scanning is performed. A specific pattern (striped pattern) generated in the line direction can be executed at high speed with a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required, and the quality of a face image is determined. In addition, each image data can be determined by one frame of image data, not only between a series of image data captured continuously but also between discontinuous image data captured at random timing. Judgment can also be executed reliably.

  In addition, here, detection of a predetermined part and detection of a specific pattern in the predetermined part are performed by a filter of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the interlace method of the image sensor. Therefore, while further reducing the circuit scale and the specific pattern (striped pattern) generated in the scanning line direction of the interlace method, it can be executed at high speed by a simple process with a small amount of calculation, When determining the quality of a face image without requiring continuous image data of a plurality of frames, determination can be made with one frame of image data for each image data. In addition to this determination, it is possible to reliably execute determination of discontinuous image data captured at random timing.

  (13) In the invention according to claim 13, in the face image determination device according to any one of (7) to (11), a detection result of a specific pattern in a predetermined part of the image data of the face image is set to a predetermined threshold value. And the image data whose signal value change is not more than the threshold value is determined to be good.

  In other words, the quality of the face image is determined by comparing the signal value, which is the detection result of a specific pattern at a predetermined part in each image data, with a predetermined threshold value. When determining the quality of a face image without requiring data, it is possible to determine with one frame of image data for each image data, not only for determination of a series of image data taken continuously, It is also possible to determine discontinuous image data captured at random timing.

  (14) In the invention described in claim 14, in the face image determination device according to any one of (7) to (12), detection of a specific pattern in a predetermined part is executed for a plurality of stored image data. The determination unit executes determination according to the detection results for the plurality of image data, and determines whether each image data can be output according to the determination result of the plurality of image data.

  In other words, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and whether or not each image data can be output is determined according to the quality of the determination results of the plurality of image data. Therefore, it is possible to determine the quality of a face image without using multiple frames of continuous image data, and it is possible to determine with one frame of image data for each image data. In addition to the determination of a series of image data, it is possible to determine discontinuous image data captured at random timing, and it is also possible to select which of a plurality of image data is to be output.

  (15) The invention according to claim 15 is the face image capturing device according to any one of (8) to (14), wherein a series of a plurality of image data is stored, and any of the stored series of image data When the specific pattern is detected, the pixel position where the specific pattern is generated is created as a pattern distribution, and the other image data in the series of image data is indicated by the pattern distribution. Pattern detection is performed at the pixel position.

  That is, since the position where the blink occurs in the image data in which the blink is detected is obtained as a pattern distribution, and the pattern detection is performed on the remaining image data of the series of image data with respect to the pattern distribution, Of the series of images at the timing before and after the blink, those that do not detect the blink are determined as the high-quality images. Therefore, it is possible to reliably detect the high-quality images that do not include the blink. Further, the moment when the subject blinks is detected and identified, the region of both eyes is identified by the distribution of the specific pattern of the blinked image data, and other image data is detected for the identified region. Therefore, it is possible to more reliably detect a high-quality image that does not include blinking.

  (16) In the invention of the face image determination program according to the sixteenth aspect, image data is generated by taking an image with an image sensor driven by an interlace method, and a specific part of the image data of the face image is specified at a predetermined part. A pattern is detected, and the quality of the face image is determined based on the detection result of the specific pattern.

  The predetermined part is an eye part of the face image. The specific pattern is a striped pattern appearing in a frame image composed of a field image with eyes open and a field image with eyes closed. Alternatively, the specific pattern is a striped pattern appearing in a frame image formed by field images having different eye opening degrees. Here, the stripe pattern occurs in the scanning line direction of the interlace method.

  That is, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and therefore the quality of the face image is determined without requiring continuous image data of a plurality of frames. In this case, each image data can be determined by one frame of image data, and not only a series of image data continuously captured but also discontinuous image data captured at random timings. It is also possible to determine.

The best mode for carrying out the present invention will be described below in detail with reference to the drawings.
A preferred embodiment of a face image photographing device, a face image determining device, and a face image determining program according to the best mode for carrying out the present invention will be described. This does not limit the scope of the present invention.

<First Embodiment>
overall structure:
A face image photographing apparatus 100 shown in FIG. 1 constitutes a main part of a face image photographing system for photographing a face photograph used for various identification cards and the like. Here, a video signal or image data generated by an imaging unit having an imaging element that is driven by 2: 1 interlace, such as a television system, is used. In the face image photographing apparatus 100, it is important that the photographed face image is not a moment such as blinking. For this reason, it has the function to determine whether the face image image | photographed with the face image imaging device 100 has blinked.

  Here, the first embodiment of the present invention will be described using the face image capturing device 100 as a specific example, but the control program that operates inside the face image capturing device 100 is a face image determination program. .

  Also, here, the first embodiment of the present invention will be described using the face image photographing device 100 as a specific example, but the photographing unit is independent and receives video signals and image data from an external photographing unit. The face image determination device that performs determination such as blinking also performs the same operation.

  Here, as shown in FIG. 1, a face image photographing apparatus 100 that photographs a face image of a subject 200 includes, as main components, a control unit 101, an operation unit 102, a display unit 104, an I / O 106, and an illumination unit. 110, an imaging unit 120, a storage unit 130, and an image processing unit 140. Here, the image processing unit 140 includes a part recognition unit 150, a pattern detection unit 160, a determination unit 170, and a selection unit 180. In FIG. 1, known circuit configurations such as a power supply unit and various buses are omitted.

  The control unit 101 is configured by a CPU (Central Processing Unit), a ROM (Read Only Memory), and the like (not shown), and performs face image shooting and face image determination (blink) in cooperation with various programs stored in the ROM. Various types of control and various types of processing are executed.

  The operation unit 102 includes various function buttons, various knobs, icons on a display unit described later, and the like, and outputs an operation signal corresponding to a user operation to the control unit 101. Alternatively, the control unit 101 reads the operation state of the user on the operation unit 102.

  The display unit 104 is configured by a flat panel display such as a liquid crystal display (LCD) or an organic EL (ElectroLuminescence), and displays various states based on instructions from the control unit 101, and the operations performed by the operation unit 102 described above. Assist display for input, display of captured face image, etc.

  The I / O 106 converts a control signal from the control unit 101 into a signal format suitable for the illumination unit 110 and the imaging unit 120 and transmits the signal, and transmits a video signal or image data from the imaging unit 120 to the control unit 101. .

  The illumination unit 110 is an illumination unit that receives control of the control unit 101 via the I / O 106 and illuminates the subject 200 with appropriate exposure centered on the face of the subject 200 using a strobe light or the like at the timing of a trigger signal at the time of shooting. is there.

  The image capturing unit 120 is an image capturing unit having an image sensor that is driven by 2: 1 interlace, such as a television system, and receives a trigger signal for image capturing from the control unit 101 to generate a video signal from two fields. One frame of image data is generated, and this image data is sent to the control unit 101.

  The storage unit 130 includes a volatile memory such as SRAM (Static Random Access Memory) or SDRAM (Synchronous Dynamic Random Access Memory), or a non-volatile memory such as a flash memory. The storage unit 130 stores various setting data of the face image photographing device 100, photographed image data, and the like.

  The image processing unit 140 uses a specific pattern (hereinafter, referred to as a predetermined pattern) in a predetermined part of the image data of the face image generated by imaging with the imaging unit 120 having an imaging device that is driven in an interlaced manner. A specific pattern) is detected, and the quality of the face image is determined based on the detection result of the specific pattern.

  The image processing unit 140 has the following functions (part recognition, specific pattern detection, image pass / fail determination, output availability selection), and can also be configured by the control unit 101 and an image processing program. It is also possible to use a dedicated image processor or image processing circuit.

  The part recognition unit 150 recognizes a predetermined part of the image data of the face image in which a specific pattern described later appears. Here, the predetermined part is an eye part in the face image. This part recognition can be performed by recognition by pattern matching of both eyes after recognition of the face outline by pattern matching. However, since it is sufficient that at least a part can be recognized when a specific pattern to be described later is generated, the predetermined part is provided by a filter of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the interlace method of the image sensor. Can be detected. By using such a filter, it can be executed at high speed by simple processing with a small amount of calculation.

  The pattern detection unit 160 detects a specific pattern in a predetermined part of the image data of the face image. Here, the predetermined part is an eye part of the face image. Further, the specific pattern is a fringe appearing in a frame image (FIG. 2C) of a field image (FIG. 2A) with eyes open and a field image with eyes closed (FIG. 2B). It is a pattern of patterns. Alternatively, the specific pattern is a striped pattern appearing in a frame image (FIG. 2C) based on field images (FIGS. 2A and 2B) having different eye opening degrees. Here, the stripe pattern is generated in the interlaced scanning line direction, such as a white eye and an eyelid portion, and a black eye and an eyelid portion.

  The pattern detection unit 160 can detect a specific pattern in a predetermined region by using a filter 161 (see FIG. 3) of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the interlace method of the image sensor. Is possible. By using such a filter 161, a specific pattern (striped pattern) generated in the scanning line direction of the interlace method can be executed at high speed by a simple process with a small amount of calculation, and a continuous image of a plurality of frames is obtained. When determining the quality of a face image without requiring data, it is possible to determine with one frame of image data for each image data, not only for determination of a series of image data taken continuously, Determination of discontinuous image data captured at random timing can also be reliably performed.

  FIG. 3 is an explanatory diagram of the filter 161 having at least three pixels (five pixels in FIG. 3) in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor described above. Here, a specific example of a pixel having a total of five elements having a3 at the center as a target pixel, a2 and a1 on the upper side, and a4 and a5 on the lower side is shown. The specific pattern to be detected is as shown in FIG. 2 (c), and each pixel is white / black or bright / dark. Therefore, at least three pixels are used for the interlace method of the image sensor. It can be detected by a one-dimensional filter in a direction orthogonal to the drive direction.

  The determination unit 170 determines the quality of the face image according to the presence / absence of the specific pattern in the pattern detection unit 160 described above, the intensity of the specific pattern (horizontal stripe light / dark difference), the blink level, and the like. Here, the intensity of the specific pattern means the absolute value C of the difference between the target pixels in the filter processing by the filter 161 described above. Further, in each image data, the horizontal stripe area pixel total number D, the horizontal stripe light / dark difference E (= the sum of the light and dark differences of the detected pixels and the horizontal stripe pixel total number D), and the blink F (= horizontal stripe light / dark difference E × The total number of horizontal stripe pixels D) is obtained.

  The selection unit 180 determines whether to output each image data according to the determination result of the determination unit 170 described above, determines the best image in a plurality of image data according to the determination result of the determination unit 170 described above, and determines whether output is possible. And so on. Note that the determination by the determination unit 170 and the selection by the selection unit 180 are not only determination and selection of a series of image data captured continuously, but also discontinuous images captured at random timing. It is also possible to determine and select between data.

Operation:
Hereinafter, the operation of the face image capturing device 100 of the first embodiment will be described with reference to the flowchart shown in FIG.

  Here, the control unit 101 monitors an operation input to the operation unit 102, and in response to an operation of a button or icon corresponding to a release button on the operation unit 102, the control unit 101 communicates with the illumination unit 110 via the I / O 106. A trigger signal for photographing is supplied to the imaging unit 120.

  In response to the trigger signal, the illumination unit 110 illuminates the face portion of the subject 200. Further, the imaging unit 120 performs imaging in response to the trigger signal, and generates image data of one frame from a video signal of two fields by an imaging device that is driven by interlace. The image data is transmitted via the I / O 106. To the control unit 101. Then, the storage unit 130 stores the image data generated by the imaging unit 120 in response to the trigger signal in a predetermined storage area (S1 in FIG. 4). Note that the generation of image data by photographing with the imaging unit 120 and the storage of image data in the storage unit 130 are determined in advance for a predetermined number of the same subject in consideration of generation of a defective image due to blinking or the like. (For example, 10 sheets) are repeated (FIG. 4, S2, S1).

  Then, when the above-described generation of image data by photographing with the imaging unit 120 and storage of image data in the storage unit 130 are repeated a predetermined number of times (for example, 10), the following face image pass / fail judgment is performed. That is, a determination is made by detecting the presence or absence of blinks. First, the control unit 101 specifies image data for which pass / fail judgment is performed from among a plurality of stored image data (S3 in FIG. 4). Here, the control unit 101 first specifies the first of the ten pieces of image data as the object of the quality determination.

Here, with respect to the image data specified as the pass / fail judgment target, the part recognition unit 150 performs a predetermined part (face) of the image data of the face image in which the specific pattern shown in FIG. Prior to the pattern detection, the regions of both eyes in the image are recognized (S4 in FIG. 4).

  The recognition of the predetermined part in the part recognition unit 150 can be performed by the recognition by the pattern matching of both eyes after the recognition of the face contour by the pattern matching. However, since it is sufficient that at least a part can be recognized when a specific pattern to be described later is generated, a predetermined part is obtained by a filter of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor. Can be detected. By using such a filter, it can be executed at high speed by simple processing with a small amount of calculation.

  After the above predetermined part recognition is completed, the pattern detection unit 160 performs filtering processing on the target pixel designated by the instruction of the control unit 101 (S5 in FIG. 4) in the part recognized as the binocular part by the part recognition. A horizontal stripe pattern (FIG. 2C) which is a specific pattern is detected by (FIG. 4S6) (FIG. 4S7). Here, the target pixel is designated by the control unit 101 in order from the pixel with the smallest address within the range of the part recognized as the binocular part.

  In this pattern detection unit 160, the filter processing of at least three pixels in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor (see FIG. 3) is performed with the above target pixel as a center. Execute.

  Here, when the filter 161 is a five-element filter a1 to a5 shown in FIG. 3, assuming that the signal values of the respective elements are A1 to A5, A1-A2> B, A3-A2> B, and If A3-A4> B and A5-A4> B, then horizontal stripes of light / dark / light / dark / light are present. On the other hand, when the filter 161 is a five-element filter of A1 to A5, A2-A1> B, A2-A3> B, A4-A3> B, and A4-A5> B. For example, horizontal stripes of dark, light, dark, light, and dark exist. Here, B may be 0, but a different value may be determined depending on the noise component included in the image data.

  Here, when the specific pattern is detected by the pattern detection unit 160 described above in the above target pixel, the determination unit 170 first calculates the intensity (horizontal stripe light / dark difference) of the specific pattern (S8 in FIG. 4).

  Here, the intensity of the specific pattern means the absolute value C of the difference between the target pixels in the filter processing by the filter 161 described above, and C = {(A1−A2) − (A2−A3) + (A3). -A4)-(A4-A5)} / 4. In each image data, the initial value of the total number D of horizontal stripe area pixels is set to 0, and D is incremented by 1 each time the target pixel is detected as a horizontal stripe area by the filter 161 described above. In addition, the determination unit 170 sets the horizontal stripe light / dark difference E as E = total sum of light / dark differences of pixels detected as a specific pattern / horizontal stripe pixel total number D, and the blink degree F as F = horizontal stripe light / dark difference E × horizontal stripe total pixel number. D, and so on.

  When the determination unit 170 completes the calculation of the absolute value C and the total number D of pixels for the target pixel, the control unit 101 assigns the target pixel to the adjacent pixel in the recognized region (S10 in FIG. 4). The above filter processing (FIG. 4S6), detection of a specific pattern (FIG. 4S7), calculation of absolute value C of difference and total number of pixels D, calculation of horizontal stripe light / dark difference E, and calculation of blinking degree F (FIG. 4S8), The process is continued until all pixels in the recognized part are completed (S9 in FIG. 4).

  Further, as described above, the filtering process (S6 in FIG. 4), the detection of the specific pattern (S7 in FIG. 4), the calculation of the absolute value C of the difference and the total number D of pixels and the horizontal stripes for all the pixels in the recognized part in the image. When the calculation of the light / dark difference E and the calculation of the blinking degree F (S8 in FIG. 4) are completed for the first image data, the control unit 101 next determines pass / fail from the plurality of stored image data. The image data to be performed is specified (S12 and S3 in FIG. 4). That is, in accordance with an instruction from the control unit 101, the same pass / fail determination is executed for all the predetermined plural sheets (for example, 10 sheets) of image data stored in the storage unit 130.

  For all image data of a series of images stored in the storage unit 130, pass / fail judgment (filter processing (FIG. 4S6), detection of specific pattern (FIG. 4S7) for all pixels in a recognized part in the image), When the calculation of C to F (S8 in FIG. 4) is completed, the selection unit 180 determines whether to output each image data according to the determination result in the determination unit 170 described above based on an instruction from the control unit 101. Alternatively, the best image in the plurality of image data is determined according to the determination result in the determination unit 170 described above (S13 in FIG. 4).

  In this case, the selection unit 180 or the control unit 101 assigns an output enabled flag to the image data with the image data having the minimum blink F as the best image data (best image data). In addition, the selection unit 180 or the control unit 101 assigns an output enable flag to the image data, with the image data having the blink degree F equal to or less than a predetermined threshold as image data with good quality (good quality image data). To do. Further, the selection unit 180 or the control unit 101 assigns, to the image data, a non-outputable flag, with image data having a blink degree F larger than a predetermined threshold as defective image data.

Effects obtained by the first embodiment:
According to the first embodiment of the operation as described above, the effects listed below can be obtained.

  (1-1) In this embodiment, image data is generated by imaging with an image sensor driven by an interlace method, a specific pattern in a predetermined part of the image data of the face image is detected, and this specification is performed. The quality of the face image is determined based on the pattern detection result. As a result, the quality determination of the face image is executed based on the detection of the specific pattern at a predetermined part in each image data. Therefore, the quality determination of the face image is performed without requiring continuous image data of a plurality of frames. In this case, each image data can be determined by one frame of image data, and not only a series of image data continuously captured but also discontinuous image data captured at random timings. It is also possible to determine.

  (1-2) In this embodiment, in the face image capturing device of (1-1), when recognizing a predetermined part of a face image, 1 in a direction orthogonal to the driving direction of the interlace method of the image sensor A filter in a dimensional direction is provided, and a predetermined part is detected by a change in signal value by the filter. That is, since a predetermined part is detected by a one-dimensional filter in a direction orthogonal to the interlaced driving direction of the image sensor, an interlaced scanning line is applied to the eye part of the face image. A specific pattern (striped pattern) that occurs in the direction can be executed at high speed by a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required. Each image data can be determined by one frame of image data, and not only a series of image data continuously shot but also a discontinuous image data shot at random timing can be determined. It becomes possible to execute reliably.

  (1-3) In this embodiment, in the face image capturing device of (1-1) or (1-2), when detecting a specific pattern in a predetermined part, the interlace driving direction of the image sensor Includes a one-dimensional filter in an orthogonal direction, and detects a specific pattern at a predetermined portion by a change in signal value using the filter.

  (1-3-1) Here, since a specific pattern is detected in a predetermined part by a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor, the interlace method is used. The specific pattern (striped pattern) generated in the scanning line direction can be executed at high speed by a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required, and the quality of the face image is determined. In this case, each image data can be determined by one frame of image data, and not only a series of image data continuously captured but also discontinuous image data captured at random timings. It is possible to reliably execute the determination.

  (1-3-2) Here, a predetermined part is detected and a specific pattern is detected in the predetermined part by a one-dimensional filter in a direction orthogonal to the interlaced drive direction of the image sensor. As a result, the specific pattern (striped pattern) generated in the interlaced scanning line direction can be executed at high speed with a simple processing with a small amount of computation while further reducing the circuit scale and reduction. When performing pass / fail judgment of a face image without the need for continuous image data of a plurality of frames, it is possible to make a judgment with one frame of image data for each image data, and a series of image data taken continuously. In addition to the determination of each other, it is possible to reliably execute determination of discontinuous image data captured at random timing.

  In the embodiment of (1-4), in the face image capturing device of (1-3), the signal value is obtained by a filter of at least three pixels in a one-dimensional direction orthogonal to the interlaced drive direction of the image sensor. A predetermined part is detected by the change.

  (1-4-1) Here, a specific pattern in a predetermined part is detected by a filter of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the interlace method of the image sensor. A specific pattern (striped pattern) generated in the scanning line direction of the interlace method can be executed at high speed by a simple process with a small amount of calculation, and continuous image data of a plurality of frames is not required. When performing pass / fail judgment, it is possible to make a judgment with one frame of image data for each image data, and not only a series of consecutively photographed image data, but also a discontinuity photographed at random timing Therefore, it is possible to reliably execute determination between image data.

  (1-4-2) Here, detection of a predetermined part and a specific pattern in the predetermined part by a filter of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the interlace method of the image sensor Detect specific patterns (stripes) generated in the interlaced scanning line direction at a high speed with a simple processing with a small amount of computation, while further reducing the circuit scale and reduction. It is possible to determine the quality of a face image without the need for continuous image data of a plurality of frames, and it is possible to determine with one frame of image data for each image data. In addition to the determination of image data, it is possible to reliably execute determination of discontinuous image data captured at random timing.

  (1-5) In this embodiment, in the face image capturing device of any one of (1-1) to (1-4), the detection result of the specific pattern in a predetermined part of the image data of the face image is predetermined. Compared with the threshold value, image data whose signal value change is not more than a threshold value is determined to be good. In other words, the quality of the face image is determined by comparing the signal value, which is the detection result of the specific pattern at a predetermined part in each image data, with a predetermined threshold value. When determining the quality of a face image, it is possible to determine with one frame of image data for each image data, and not only for determination of a series of image data taken continuously but also for random It is also possible to determine discontinuous image data captured at various timings.

  (1-6) In this embodiment, in the face image capturing device of any one of (1-1) to (1-5), a specific pattern is detected at a predetermined site for a plurality of stored image data. The determination unit executes determination according to detection results for a plurality of image data, and determines whether each image data can be output according to whether the determination results for the plurality of image data are acceptable. That is, the quality of the face image is determined based on the detection of a specific pattern at a predetermined part in each image data, and the output of each image data is determined according to the quality of the determination results of the plurality of image data. Therefore, when determining the quality of a face image without the need for continuous image data of a plurality of frames, it is possible to determine with one frame of image data for each image data, and a series of images taken continuously. In addition to the determination of the image data, discontinuous image data captured at random timing can be determined, and it is possible to select which of the plurality of image data is to be output.

<Second Embodiment>
The second embodiment is realized by a face image photographing device 300 shown in FIG. That is, it is a face image photographing device 300 that constitutes a main part of a face image photographing system for photographing a face photograph used for various identification cards and the like, and is driven by a 2: 1 interlace such as a television system. Image data generated by an imaging unit having an imaging element is used.

  In the face image photographing apparatus 300, it is important that the photographed face image is not a moment such as a blink. For this reason, it has the function to determine whether the face image image | photographed with the face image imaging device 300 is blinking.

  Here, the second embodiment of the present invention will be described using the face image capturing device 300 of FIG. 5 as a specific example, but a control program that operates inside the face image capturing device 300 performs face image determination. This corresponds to the second embodiment of the program. Here, the second embodiment of the present invention will be described using the face image capturing device 300 as a specific example, but the image capturing unit is independent and receives video signals and image data from an external image capturing unit. The face image determination device that performs determination such as blinking also performs the same operation.

  In the second embodiment, as will be described later, the moment when the subject blinks is detected and specified, and the image that does not detect blinking is determined as a high-quality image among a series of images at timings before and after that. It is a feature.

overall structure:
Here, as shown in FIG. 5, a face image photographing apparatus 300 that photographs a face image of the subject 200 includes, as main components, a control unit 301, an operation unit 302, a display unit 304, an I / O 306, and an illumination unit. 310, an imaging unit 320, a storage unit 330, and an image processing unit 340. Here, the image processing unit 340 includes a pattern detection unit 360, a pattern distribution creation unit 365, a determination unit 370, and a selection unit 380. In FIG. 5, known circuit configurations such as a power supply unit and various buses are omitted.

  The control unit 301 is configured by a CPU (Central Processing Unit), a ROM (Read Only Memory), and the like (not shown), and performs face image shooting and face image determination (blink) in cooperation with various programs stored in the ROM. Various types of control and various types of processing are executed.

  The operation unit 302 includes various function buttons, various knobs, icons on a display unit to be described later, and the like, and outputs an operation signal corresponding to a user operation to the control unit 301. Alternatively, the control unit 301 reads the operation state of the user on the operation unit 302.

  The display unit 304 is configured by a flat panel display such as a liquid crystal display (LCD) or an organic EL (ElectroLuminescence), and displays various states based on instructions from the control unit 301, and the operations performed by the operation unit 302 described above. Assist display for input, display of captured face image, etc.

  The I / O 306 converts the control signal from the control unit 301 into a signal format suitable for the illumination unit 310 and the imaging unit 320 and transmits it, and transmits the video signal or image data from the imaging unit 320 to the control unit 301. .

  The illumination unit 310 is an illumination unit that receives control of the control unit 301 via the I / O 306 and illuminates the subject 200 with appropriate exposure centered on the face of the subject 200 using a strobe light or the like at the timing of a trigger signal at the time of shooting. is there.

  The image pickup unit 320 is an image pickup unit having an image pickup element that is driven by 2: 1 interlace, such as a television system, and receives a trigger signal for shooting from the control unit 301 and generates a video signal from two fields. One frame of image data is generated, and this image data is sent to the control unit 301.

  The storage unit 330 is configured by a volatile memory such as SRAM (Static Random Access Memory) or SDRAM (Synchronous Dynamic Random Access Memory), or a non-volatile memory such as a flash memory. The storage unit 330 stores various setting data of the face image photographing device 300, photographed image data, and the like.

  The image processing unit 340 detects a specific pattern at a predetermined part of the image data of the face image generated by imaging with the imaging unit 320 having an imaging element that is driven in an interlaced manner, The quality of the face image is determined based on the detection result of the specific pattern.

  The image processing unit 340 has the following functions (specific pattern detection, pattern distribution creation, image quality determination, output availability selection), and can also be configured by the control unit 301 and an image processing program. It is also possible to use a dedicated image processor or image processing circuit.

  The pattern detection unit 360 detects a specific pattern in a predetermined part of the image data of the face image. Here, the predetermined part is the part of the eyes (both eyes) of the face image. Further, the specific pattern is a fringe appearing in a frame image (FIG. 2C) of a field image (FIG. 2A) with eyes open and a field image with eyes closed (FIG. 2B). It is a pattern of patterns. Alternatively, the specific pattern is a striped pattern appearing in a frame image (FIG. 2C) based on field images (FIGS. 2A and 2B) having different eye opening degrees. Here, the stripe pattern is generated in the interlaced scanning line direction, such as a white eye and an eyelid portion, and a black eye and an eyelid portion.

  The pattern detection unit 360 can detect a specific pattern at a predetermined portion by a filter 161 (see FIG. 3) of at least three pixels in a one-dimensional direction orthogonal to the driving direction of the image sensor interlace method. Is possible. By using such a filter 161, a specific pattern (striped pattern) generated in the scanning line direction of the interlace method can be executed at high speed by a simple process with a small amount of calculation, and a continuous image of a plurality of frames is obtained. When determining the quality of a face image without requiring data, it is possible to determine with one frame of image data for each image data, not only for determination of a series of image data taken continuously, Determination of discontinuous image data captured at random timing can also be reliably performed.

  FIG. 3 is an explanatory diagram of the filter 161 having at least 3 pixels (5 pixels in FIG. 3) in a one-dimensional direction perpendicular to the driving direction of the interlace method of the image sensor described above. Here, a specific example of a pixel having a total of five elements having a3 at the center as a target pixel, a2 and a1 on the upper side, and a4 and a5 on the lower side is shown. The specific pattern to be detected is as shown in FIG. 2 (c), and each pixel is white / black or bright / dark. Therefore, at least three pixels are used for the interlace method of the image sensor. It can be detected by a one-dimensional filter in a direction orthogonal to the drive direction.

  The pattern distribution creation unit 365 refers to the detection result of the pattern detection unit 360 described above, and generates a pattern distribution of the part of the image data of the face image in which the specific pattern appears. Here, the part where the specific pattern appears is a part of both eyes in the face image. That is, the position data of the pixel of the specific pattern included in the image data is generated as a pattern distribution.

  The determination unit 370 relates to the pixel position indicated in the pattern distribution created from any one of the image data by the pattern distribution creation unit 365 described above, the intensity of a specific pattern (horizontal stripe light / dark difference) in other image data, the blinking degree, etc. To determine the quality of the face image.

  Here, the intensity of the specific pattern means the absolute value C of the difference between the target pixels in the filter processing by the filter 161 described above. Further, in each image data, the horizontal stripe area pixel total number D, the horizontal stripe light / dark difference E (= the sum of the light and dark differences of the detected pixels and the horizontal stripe pixel total number D), and the blink F (= horizontal stripe light / dark difference E × The total number of horizontal stripe pixels D) is obtained.

  In the second embodiment, the blinking degree F is determined for a part of a predetermined number of series of image data where the distribution of the specific pattern by this filtering process exists, and for the other parts of the image data. It is characterized by performing.

  The selection unit 380 determines whether to output each image data in accordance with the determination result in the determination unit 370 described above, and determines a high-quality image in the plurality of image data in accordance with the determination result in the determination unit 370 described above.

Operation:
Hereinafter, the operation of the face image capturing device 300 of the second embodiment will be described with reference to the flowchart shown in FIG.

  Here, the control unit 301 monitors an operation input to the operation unit 302, and the lighting unit 310 is connected via the I / O 306 according to the operation of a button or icon corresponding to the release button on the operation unit 302. A trigger signal for photographing is supplied to the imaging unit 320.

  In response to the trigger signal, the illumination unit 310 illuminates the face portion of the subject 200. In addition, the imaging unit 320 performs imaging in response to the trigger signal, and generates one frame of image data from a two-field video signal by an imaging device that is driven by interlace. The image data is transmitted via the I / O 306. To the control unit 301. Then, the storage unit 330 stores the image data generated by the imaging unit 320 in accordance with the trigger signal in a predetermined storage area.

  Here, the generation of image data by photographing with the imaging unit 320 and the storage of image data in the storage unit 330 are determined in advance for the same subject in consideration of generation of defective images due to blinking or the like. The number of sheets (for example, 10 sheets) is repeated (S1 in FIG. 6).

  The second embodiment is characterized in that the moment when the subject blinks is detected and specified, and a series of images at the timing before and after that is determined as a good quality image in which no blink is detected. For this reason, a plurality of predetermined images are taken within a certain time for the same subject, but the plurality of image data need not be continuous frame images at a frame rate (such as every 1/30 seconds). There is no problem even if there is a series of image data taken at a discontinuous timing such as every 0.5 seconds or every 1 second.

  Here, the control unit 301 gives an instruction to the pattern detection unit 360 and the pattern distribution creation unit 365, executes a filtering process on any of a predetermined number of series of image data (S4 in FIG. 6), and identifies the target pixel. If it is a pattern (Y in FIG. 6 S), the coordinates (position data) of the pixel of interest and the filter result are stored in the storage unit 330 (FIG. 6 S6), and the above processing is executed for the entire area of the image data (S3 in FIG. 6). Y), a specific pattern distribution is created (S7 in FIG. 6).

  Then, the determination unit 370 clusters the distribution of the above specific patterns (S8 in FIG. 6), and determines whether there are two approximate areas in the same image in order to determine whether the eye is binocular (S9 in FIG. 6). ). If there is no distribution of two specific patterns approximated in the image data (N in FIG. 6 S9), the same processing is executed for the other image data.

  If the distribution of the two specific patterns does not exist in all of the predetermined number of series of image data (N in FIG. 6 S10), the imaging unit 320 newly sets a predetermined number of series of images under the control of the control unit 301. Photographing is executed and stored in the storage unit 330 (S1 in FIG. 6).

  Here, when there is no distribution of two specific patterns in all of a predetermined number of series of image data (N in FIG. 6), there is a possibility that all of the series of image data is a good quality image. Since there are various possibilities such as pattern detection failure, detection of blinking with only one eye, and no face in the vicinity of the center of the image data, a series of image data is discarded as a precaution, and the imaging unit 320 Then, a predetermined number of series of images are taken and stored in the storage unit 330 (S1 in FIG. 6).

  Further, if there is a distribution of two specific patterns in any of a predetermined number of series of image data (Y in FIG. 6 S10), the distribution of the two specific patterns blinks under the control of the control unit 301. Since it is an image at a certain moment, the determination unit 370 calculates the blinking degree F for image data other than this image (S11 in FIG. 6).

  Further, the determination unit 370 compares the blink degree F calculated as described above with a predetermined threshold value (S12 in FIG. 6). If the blink degree F is not less than the threshold value (N in FIG. 6 S12), the determination unit 370 similarly calculates the blink degree F for other image data in the series of image data (S11 in FIG. 6). And the comparison (S12 in FIG. 6).

  Here, if the blink level F is not less than the threshold value in all the image data to be searched in the series of image data (N in FIG. 6 S12), it means that all the image data has been shot in the blink state. As a result, the control unit 301 controls to discard a series of image data whose blink degree F does not satisfy the condition, and at the same time, the imaging unit 320 newly captures a predetermined number of images and stores them. The data is stored in the unit 330 (S1 in FIG. 6).

  Also, here, if the blink degree F is less than the threshold value in any one of the series of image data (N in FIG. 6 S12), it means that the image is not blinked. Therefore, the control unit 301 determines that the image data is a high-quality image (S13 in FIG. 6), adds a flag indicating that the image is a high-quality image to the image data, and ends this processing. Note that the same calculation and comparison of the blinking degree F may be executed for the remaining series of image data to obtain the best image with the smallest blinking degree F.

Effects obtained by the second embodiment:
According to the second embodiment of the operation as described above, image data is generated by imaging with an image sensor that is driven in an interlaced manner, and a specific pattern in a predetermined part of the image data of the face image is detected. In addition to the effects (see the first embodiment) obtained by determining the quality of the face image based on the detection result of the specific pattern, the effects listed below can be obtained.

  (2-1) In the second embodiment, in order to detect / specify the moment when the subject blinks, and to determine that no blink is detected as a good quality image among a series of images before and after that, It becomes possible to reliably detect a high-quality image that is not included.

  (2-2) In the second embodiment, the moment when the subject blinks is detected and specified, and the region of both eyes is specified by the distribution of the specific pattern of the blinking image data. Since the detection and determination of other image data are performed for the part, a high-quality image that does not include blinking can be detected more reliably.

<Other embodiments>
In the specific example of the first embodiment described above, the face image capturing device 100 is described as a specific example. However, the image capturing unit is independent and receives a video signal or image data from an external image capturing device or an image storage device. Even with a face image determination device or a face image determination program that performs determination such as blinking, it is possible to obtain the same effect as in the case of the first embodiment described above.

  In the specific example of the second embodiment described above, the face image capturing device 300 is described as a specific example. However, the image capturing unit is independent and receives video signals and image data from an external image capturing unit or an image storage device. Even with a face image determination device or a face image determination program that performs determination such as blinking, it is possible to obtain the same effect as in the case of the second embodiment described above.

1 is a block diagram illustrating a schematic configuration of a face image photographing device according to a first embodiment of the present invention. It is explanatory drawing which shows the mode of generation | occurrence | production of the specific pattern used for the blink detection in embodiment of this invention. It is explanatory drawing which shows the filter used when detecting the specific pattern used for the blink detection in embodiment of this invention. It is a flowchart which shows the operation state of the face image imaging device of 1st Embodiment of this invention. It is a block diagram which shows schematic structure of the face image imaging device of 2nd Embodiment of this invention. It is a flowchart which shows the operation state of the face image imaging device of 2nd Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Face imaging device 101 Control part 110 Illumination part 120 Imaging part 130 Storage part 140 Image processing part 150 Part | part detection part 160 Pattern detection part 170 Judgment part 180 Selection part

Claims (16)

An imaging unit that generates and outputs image data by imaging with an image sensor that is driven in an interlaced manner;
A pattern detection unit for detecting a specific pattern in a predetermined part of the image data of the face image generated by the imaging unit;
A determination unit that determines the quality of the face image based on the detection result in the pattern detection unit;
A face image photographing apparatus. A part recognition unit for recognizing a predetermined part of the face image;
The part recognition unit includes a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor, and detects the predetermined part by a change in a signal value by the filter.
The face image photographing device according to claim 1. The pattern detection unit includes a one-dimensional filter in a direction orthogonal to a driving direction of the interlace method of the image sensor, and detects a specific pattern in the predetermined part by a change in a signal value by the filter.
The face image photographing device according to claim 1, wherein the face image photographing device is provided. The filter includes a filter of at least three pixels in a one-dimensional direction in a direction orthogonal to the driving direction of the interlace method of the imaging element.
The face image photographing device according to claim 3. The determination unit compares the detection result of the pattern detection unit with a predetermined threshold value, and determines that image data whose signal value change is equal to or less than the threshold value is good.
The face image photographing device according to any one of claims 1 to 4, wherein the face image photographing device is characterized in that: Selecting means for determining whether or not the image data can be output according to whether the determination result in the determination unit is good or bad;
In the selection unit, the detection is performed on the plurality of stored image data by the pattern detection unit, and the determination unit performs the determination according to the detection result on the plurality of image data by the pattern detection unit. Is executed, it is determined whether to output each image data according to the determination result of the plurality of image data.
The face image photographing device according to claim 1, wherein A storage unit for storing a series of image data;
A pattern distribution creation unit that creates a pixel position where the specific pattern is generated as a pattern distribution when the specific pattern is detected in any of the image data stored in the storage unit,
The pattern detection unit performs pattern detection at pixel positions indicated by the pattern distribution with respect to other image data in the series of image data.
15. The face image determination apparatus according to claim 8, wherein A signal input unit that receives supply of image data generated by imaging a face with an image sensor that is driven in an interlaced manner;
A pattern detection unit for detecting a specific pattern in a predetermined part of the image data of the face image input to the signal input unit;
A determination unit that determines the quality of the face image based on the detection result in the pattern detection unit;
A face image determination apparatus. A signal input unit that receives supply of an interlaced video signal generated by imaging a face with an image sensor that is driven in an interlaced manner;
A frame data generation unit that generates one frame of image data from an odd field and an even field of the video signal input to the signal input unit;
A pattern detection unit for detecting a specific pattern in a predetermined part of the face image of the image data generated by the frame data generation unit;
A determination unit that determines the quality of the face image based on the detection result in the pattern detection unit;
A face image determination apparatus. A part recognition unit for recognizing a predetermined part of the face image;
The part recognition unit includes a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor, and detects the predetermined part by a change in a signal value by the filter.
10. The face image determination apparatus according to claim 8 or 9, wherein: The pattern detection unit includes a one-dimensional filter in a direction orthogonal to the driving direction of the interlace method of the image sensor, and detects the predetermined part by a change in a signal value by the filter.
11. The face image determination apparatus according to claim 8, wherein The filter includes a filter of at least three pixels in a one-dimensional direction in a direction orthogonal to the driving direction of the interlace method of the imaging element.
The face image determination device according to claim 11. The determination unit compares the detection result of the pattern detection unit with a predetermined threshold value, and determines that image data whose signal value change is equal to or less than the threshold value is good.
The face image determination apparatus according to claim 8, wherein the face image determination apparatus is a part of the face image determination apparatus. Selecting means for determining whether or not the image data can be output according to whether the determination result in the determination unit is good or bad;
In the selection unit, the detection is performed on the plurality of stored image data by the pattern detection unit, and the determination unit performs the determination according to the detection result on the plurality of image data by the pattern detection unit. Is executed, it is determined whether to output each image data according to the determination result of the plurality of image data.
The face image determination apparatus according to claim 8, wherein A storage unit for storing a series of image data;
A pattern distribution creation unit that creates a pixel position where the specific pattern is generated as a pattern distribution when the specific pattern is detected in any of the image data stored in the storage unit,
The pattern detection unit performs pattern detection at pixel positions indicated by the pattern distribution with respect to other image data in the series of image data.
15. The face image determination apparatus according to claim 8, wherein An input routine that receives supply of image data generated by imaging a face with an image sensor that is driven in an interlaced manner;
A detection routine for detecting a specific pattern in a predetermined part of the image data;
A determination routine for determining pass / fail of a face image based on a detection result in the detection routine;
A face image determination program.