JP4957030B2 - Image processing apparatus, image processing method, and program - Google Patents

Description

  The present invention relates to an image processing apparatus, an image processing method, and a program, and more particularly, to an image processing apparatus, an image processing method, and a program that can obtain an optimal image for recognition processing.

In a recognition processing system that captures a user's face with an imaging device and performs recognition processing with an image processing device, various techniques have been proposed that enable recognition processing to be performed at high speed or simply (for example, patents). References 1 and 2). Conventionally, in these imaging apparatuses, imaging is performed by automatic exposure (exposure).
JP 2005-84979 A JP 2005-141720 A

  However, in the automatic exposure, the EV value of the imaging device is determined so that the entire imaging area is equivalent to the intensity of light entering through the lens. For example, when performing facial recognition, Depending on the brightness of the background other than this area, the pixel value of the face area as the recognition target may change, and recognition may not be possible.

  The present invention has been made in view of such a situation, and makes it possible to obtain an image optimal for recognition processing.

An image processing apparatus according to an aspect of the present invention includes a reference image storage unit that stores a reference image that serves as a reference for authenticating whether or not a face included in an image captured by an imaging apparatus is a registered person, A face area detecting means for detecting a face area from the image; a histogram creating means for creating a histogram of pixel values of each pixel in the face area in the image; and each face area in the image. It is determined whether or not an error between a pixel value histogram of a pixel and a pixel value histogram of each pixel of the reference image is within a predetermined range, and the image is set to be within a predetermined range when the error is not within the predetermined range. determining means for determining an exposure amount of the imaging apparatus of the captured, so that the determined the amount of exposure, and a control means for controlling the imaging device, and the image in which the error is determined to the predetermined inside The group By comparing the reference image stored in the image storage means, the face included in the image captured by the imaging device, and an authentication means for authenticating whether the person is registered Prepare.

In the image processing apparatus according to one aspect of the present invention, a reference image serving as a reference for authenticating whether or not the face included in the image captured by the imaging apparatus is a registered person is stored in the reference image storage unit. Then, a face area is detected from the image, and a histogram of pixel values of each pixel is created in the face area in the image. Then, it is determined whether or not the error between the pixel value histogram of each pixel in the face area of the image and the pixel value histogram of each pixel of the reference image is within a predetermined range. The exposure amount of the imaging device that captured the image is determined so as to be within the predetermined range , the imaging device is controlled so that the determined exposure amount is within the predetermined range, and the reference image By comparing with the reference image stored in the storage means, it is authenticated whether or not the face included in the image captured by the imaging device is a registered person . Therefore, it is possible to obtain an image optimal for recognition processing and perform authentication processing .

The face area detection means, histogram creation means, and determination means are configured by, for example, a CPU, and the control means is, for example, wired communication conforming to IEEE1394, Ethernet (registered trademark), USB (Universal Serial Bus), or the like. It is composed of a wireless LAN (Local Area Network) such as IEEE (The Institute of Electrical and Electronic Engineers) 802.11a, 802.11b, and 802.11g, or a communication unit that performs various wireless communications such as Bluetooth. The reference image storage means is composed of, for example, a hard disk.

Reference image registration means for storing the reference image in the reference image storage means is further provided, and the control means has different exposures when the reference image registration means stores the reference image in the reference image storage means. The reference image registration unit obtains a plurality of images captured by the imaging device with different exposure amounts, and obtains one of the acquired images. A sheet can be stored in the reference image storage means as the reference image.
In the reference image registration unit, the center value of the histogram of the pixel values of the pixels of the face area of each image among the plurality of acquired images is set to the center value of the range that the pixel value can take. An image having a close and wide distribution can be stored in the reference image storage unit as the reference image.
The determining means can determine the exposure amount of the imaging apparatus by determining the shutter speed or F value of the imaging apparatus. Therefore, an image optimal for recognition processing can be obtained.

In the determination unit, the imaging is performed so that an error between a center value of a pixel value histogram of each pixel of the face area and a center value of a histogram of pixel values of each pixel of the reference image is within a predetermined range. The amount of exposure of the device can be determined. Therefore, an image optimal for recognition processing can be obtained.

According to an image processing method of one aspect of the present invention, a reference image serving as a reference for authenticating whether or not a face included in an image captured by an imaging device is a registered person is stored in a predetermined storage unit. And detecting a face area from the image, creating a histogram of pixel values of each pixel in the face area of the image, and a histogram of pixel values of each pixel of the face area of the image; the error between the histogram of pixel values of each pixel of the reference image is determined whether or not a predetermined inside, the so error is given in when not within a predetermined, of the imaging device imaging the image An exposure amount is determined, the image pickup apparatus is controlled so as to be the determined exposure amount, the image for which the error is determined to be within the predetermined range, and the reference image stored in the storage unit By comparing the imaging The face contained in the image captured by the location comprises the step of authenticating whether a person has been registered.

In a program according to one aspect of the present invention, a reference image that serves as a reference for authenticating whether a face included in an image captured by an imaging device is a registered person is stored in a predetermined storage unit, Detecting a face area from the image, creating a histogram of pixel values of each pixel in the face area of the image, and a histogram of pixel values of each pixel of the face area of the image; error between the histogram of the pixel value of each pixel of the reference image is determined whether or not a predetermined inside, the exposure amount of the error is such that the predetermined when it is not within the predetermined, the imaging device imaging the image The image pickup apparatus is controlled so that the determined exposure amount is obtained, and the image in which the error is determined to be within the predetermined range and the reference image stored in advance in the storage unit By comparing The face contained in the image captured by the device to execute processing comprising the step of authenticating whether a person is registered in the computer.

In the image processing method and program according to one aspect of the present invention, a reference image serving as a reference for authenticating whether or not the face included in the image captured by the imaging device is a registered person is stored in a predetermined storage unit. The face area is detected from the image, and a histogram of pixel values of each pixel is created in the face area in the image. Then, it is determined whether or not the error between the pixel value histogram of each pixel in the face area of the image and the pixel value histogram of each pixel of the reference image is within a predetermined range. The exposure amount of the imaging device that has captured the image is determined so as to be within the predetermined range , the imaging device is controlled so that the determined exposure amount is within the predetermined range, and the storage unit Is compared with the reference image stored in the image, it is authenticated whether or not the face included in the image captured by the imaging device is a registered person . Therefore, it is possible to obtain an image optimal for recognition processing and perform authentication processing .

  According to the present invention, an image optimal for recognition processing can be obtained.

  FIG. 1 shows a configuration example of a first embodiment of an image processing system to which the present invention is applied. In FIG. 1, a thick solid line represents image data, and a thin solid line represents control data.

  The image processing system 1 in FIG. 1 performs recognition processing of an imaging device 11 that captures a human face as a recognition target and an image including a face supplied from the imaging device 11 (hereinafter referred to as a face image). The image processing apparatus 12 is comprised.

  The imaging device 11 captures a face as a recognition target based on a timing signal representing imaging timing supplied from the image processing device 12 and supplies a face image obtained as a result to the image processing device 12. The imaging device 11 can perform imaging by automatic exposure (exposure) and imaging by manual exposure.

When imaging is performed by automatic exposure (exposure), the imaging device 11 supplies the set exposure amount to the image processing device 12. This exposure amount can be generally expressed as an EV value. The EV value is determined by an aperture value F, a shutter speed T, and an imager sensitivity ISO corresponding to film sensitivity, and is expressed by the following equation.
EV value = log ₂ F ² -Log ₂ T-log ₂ (ISO / 100)

  Since the sensitivity ISO of the imager is a fixed value for each imaging device, it is stored in advance in the image processing device 12. The imaging device 11 uses the aperture value F and the shutter speed T set at the time of imaging as the image processing device. 12 is supplied. The imager sensitivity ISO may be supplied along with the aperture value F and shutter speed T each time.

  On the other hand, when imaging is performed by manual exposure, the imaging device 11 captures images with the aperture value F and the shutter speed T supplied from the image processing device 12.

  The image processing apparatus 12 includes a frame memory 21, a face area detection unit 22, an image determination unit 23, a recognition processing unit 24, a reference image storage unit 25, and a communication unit 26.

  The frame memory 21 temporarily stores the face image supplied from the imaging device 11 and supplies it to the face area detection unit 22 and the image determination unit 23 as necessary. The face area detection unit 22 detects a face area from the face image supplied from the frame memory 21, and supplies information (face area information) representing the face area obtained as a result to the image determination unit 23. Note that the face region detection method performed by the face region detection unit 22 is not particularly limited, and any method can be employed. For example, the face region detection unit 22 detects a face region by extracting a contour portion from the supplied face image and detecting a portion that matches the contour of the face among the extracted contour portions. Further, for example, the face area detection unit 22 labels the supplied face image, and detects an image having the largest size or a size within a predetermined range among the labeled images as a face area.

  A face image is supplied from the frame memory 21 to the image determination unit 23, and face area information is supplied from the face area detection unit 22. The image determination unit 23 specifies a face region image (face region image) from the face image based on the face region information supplied from the face region detection unit 22. Then, the image determination unit 23 determines whether or not the specified face area image is an image suitable for recognition processing. Specifically, the image determination unit 23 determines an error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image that is stored in the reference image storage unit 25 and serves as a reference for recognition processing. It is determined whether or not the face area image is an image suitable for recognition processing based on whether or not is within a predetermined range. Here, the central value of the luminance distribution of the reference image or the face area image is the (maximum value−minimum value) / pixel value of the pixel value histogram represented by the pixel value on the horizontal axis and the number of pixels on the vertical axis. The value obtained in 2. Note that the center value of the luminance distribution of the reference image or the face area image may be the pixel value having the largest number of pixels or the average value of the pixel values. That is, the center value of the luminance distribution of the reference image or face area image may be any value that represents the brightness level of the entire face area image or the entire reference image. Hereinafter, when it is not necessary to distinguish the reference image and the face area image, they are referred to as recognition images.

  When it is determined that the face area image is an image suitable for recognition processing, the image determination unit 23 supplies the face area image to the recognition processing unit 24. On the other hand, when it is determined that the face area image is not an image suitable for recognition processing, the image determination unit 23 reduces an error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image. Such an aperture value F and shutter speed T are determined and supplied to the communication unit 26.

  The recognition processing unit 24 registers the face imaged by the imaging device 11 by comparing the face area image supplied from the image determination unit 23 with the reference image stored in the reference image storage unit 25. Authenticate whether you are a person.

  The reference image storage unit 25 stores a reference image that serves as a reference for authenticating whether or not the face imaged by the imaging device 11 is a registered person. The reference image is captured in advance by the imaging device 11 before the face image is captured, and stored in the reference image storage unit 25.

  The communication unit 26 controls the imaging device 11. Specifically, the communication unit 26 transmits the aperture value F and the shutter speed T from the image determination unit 23 to the imaging device 11. The communication unit 26 also transmits a timing signal to the imaging unit 11 based on a control signal from an operation unit or a control unit (not shown).

  In the image processing apparatus 12 of the image processing system 1 configured as described above, a face area is detected from the face image captured by the imaging apparatus 11. Then, the exposure amount of the imaging device 11 is controlled so that the brightness (luminance distribution) of the image of the detected face area is the same as the distribution of the reference image. As a result, an optimal face area image for recognition processing can be obtained.

  FIG. 2 shows a detailed configuration example of the imaging apparatus 11 of FIG.

  The imaging device 11 includes a lens 41, a diaphragm 42, an image sensor 43, an A / D conversion unit 44, and an imaging control unit 45.

  The light from the subject that has passed through the lens 41 and the aperture 42 enters the image sensor 43. The image sensor 43 includes, for example, a CMOS (Complementary Mental Oxide Semiconductor) sensor, a CCD (Charge Coupled Device) sensor, and the like, and images according to a timing signal from the imaging control unit 45. That is, the image sensor 43 supplies an analog electrical signal corresponding to the received light from the subject to the A / D converter 44.

  The A / D (Analog / Digital) converter 44 converts an analog electrical signal supplied from the image sensor 43 into a digital signal having a predetermined number of bits, and outputs the digital signal to the image processing device 12 as an image signal. As the A / D conversion unit 44, an A / D converter that converts an 8-bit digital signal is often used. In this case, the image output from the A / D conversion unit 44 is represented by 256 gradations.

  The imaging control unit 45 controls the timing of imaging by supplying a timing signal supplied from the image processing device 12 to the image sensor 43. Further, in the automatic exposure mode, the imaging control unit 45 controls the shutter speeds of the aperture 42 and the image sensor 43 so as to be equivalent to the illuminance of the subject measured by an exposure meter (not shown). On the other hand, in the manual exposure mode, the imaging control unit 45 sets the shutter speed of the aperture 42 and the image sensor 43 to the aperture value F and the shutter speed T supplied from the image processing device 12.

  In the imaging device 11 configured as described above, a face image including a face as a recognition target is captured by the automatic exposure or the aperture value F and the shutter speed T transmitted from the image processing device 12, and is 8-bit digital. After A / D conversion into a signal, the signal is output to the image processing device 12.

  FIG. 3 shows a detailed configuration example of the image determination unit 23 of the image processing apparatus 12.

  The image determination unit 23 includes a distribution creation unit 61, a determination unit 62, and an EV correction calculation unit 63.

  The distribution creation unit 61 of the image determination unit 23 specifies a face region portion from the face image based on the face region information supplied from the face region detection unit 22. The image of the area specified here is a face area image. In addition, the distribution creating unit 61 creates the luminance distribution of the face area image and the luminance distribution of the reference image stored in the reference image storage unit 25, and the center value of the luminance distribution of the face area image and the luminance of the reference image. Find the center value of each distribution.

  The determination unit 62 determines whether or not the face area image is an image suitable for recognition processing based on whether or not an error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image is within a predetermined range. Determine.

  If the determination unit 62 determines that the face area image is an image suitable for recognition processing, the determination unit 62 supplies the face area image to the recognition processing unit 24. On the other hand, when the determination unit 62 determines that the face area image is not an image suitable for the recognition process, the aperture value F and the shutter speed T for correcting the brightness (pixel value) of the face area image are corrected by EV. The calculation unit 63 calculates.

  The EV correction calculation unit 63 acquires the aperture value F and the shutter speed T set by the imaging device 11 when the face area image is captured from the communication unit 26. Further, when the determination unit 62 determines that the face area image is not an image suitable for recognition processing, the EV correction calculation unit 63 and the center value of the luminance distribution of the face area image and the center of the luminance distribution of the reference image The exposure amount of the imaging device 11 is determined so that the error from the value is within a predetermined range.

  Here, as described above, since the sensitivity of the imager is a fixed value, the exposure amount (EV value) is determined by the aperture value F and the shutter speed T, and the aperture value F and the shutter speed T that realize the EV value are Depending on the combination, there are several ways. For example, aperture value F = 2 and shutter speed T = 1/1000 seconds, aperture value F = 4 and shutter speed T = 1/250 seconds, aperture value F = 8 and shutter speed 1/60 seconds are the same. Exposure amount (EV value). Also, assuming that the EV value determined by automatic exposure is the reference EV value, the EV value that is twice, four times, etc. the reference EV value is the reference EV value + 1, the reference EV value + 2 The EV values whose brightness is 1/2, 1/4,... Relative to the reference EV value are the reference EV value-1, the reference EV value-2, and so on. . In the following,..., −2 EV value, −1 EV value, +1 EV value, +2 EV value,... Represent relative EV values based on the reference EV value.

  Therefore, in order to determine the exposure amount such that the error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image is within a predetermined range, the aperture value F or the shutter speed T Either one of them may be changed, or both of them may be changed. In the present embodiment, the EV correction calculation unit 63 fixes the aperture value F and changes the shutter speed T, so that the exposure amount of the imaging device 11 has an error from the central value of the luminance distribution of the reference image. It is determined to be within a predetermined range. For example, in order to double the brightness from the face image captured with the reference EV value, the shutter speed is set to ½ times without changing the aperture value F.

  The EV correction calculation unit 63 supplies the determined aperture value F and shutter speed T to the communication unit 26. Then, the aperture value F and the shutter speed T are supplied to the imaging device 11 by the communication unit 26.

  The image determination unit 23 configured as described above determines whether or not the face area image is an image suitable for recognition processing. When it is determined that the face area image is an image suitable for the recognition process, the face area image is supplied to the recognition processing unit 24. On the other hand, when it is determined that the face area image is not an image suitable for the recognition process, an aperture value F and a shutter speed T for obtaining a face area image optimal for the recognition process are obtained.

  Next, with reference to the flowchart of FIG. 4, a comparison image acquisition process for acquiring an optimal face image as a comparison image to be compared with the reference image in the recognition process will be described. This process is started when the start of the recognition process is instructed by an operation unit or a control unit (not shown) of the image processing apparatus 12.

  First, in step S <b> 1, the communication unit 26 supplies a timing signal to the imaging device 11. And the face image imaged with the imaging device 11 with the timing signal from the communication part 26 is input into the frame memory 21, and the frame memory 21 memorize | stores the face image from the imaging device 11 in step S2.

  In step S <b> 3, the face area detection unit 22 detects a face area from the face image supplied from the frame memory 21, and supplies face area information obtained as a result to the image determination unit 23.

In step S <b> 4, the distribution creation unit 61 of the image determination unit 23 creates the brightness distribution of the face image supplied from the frame memory 21 and the brightness distribution of the reference image stored in the reference image storage unit 25. In addition, the distribution creating unit 61 identifies a face region part of the luminance distribution of the face image based on the face region information supplied from the face region detecting unit 22. In step S4, the distribution creating unit 61 obtains the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image.

  In step S5, the determination unit 62 of the image determination unit 23 determines whether the error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image is within a predetermined value. 11 determines whether the face area image captured in step 11 is an image suitable for recognition processing.

  When it is determined in step S5 that the face area image currently captured by the imaging device 11 is not an image suitable for recognition processing, the EV correction calculation unit 63 in step S6 via the communication unit 26 The aperture value F and the shutter speed T when the image is captured are acquired from the imaging device 11, and the error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image is within a predetermined range. As described above, the exposure amount of the imaging device 11 is calculated and determined. In this example, a correction amount is calculated by changing only the shutter speed T out of the aperture value F and the shutter speed T supplied from the imaging device 11.

  In step S <b> 7, the communication unit 26 transmits the aperture value F and the shutter speed T supplied from the EV correction calculation unit 63 of the image determination unit 23 to the imaging device 11. Thereafter, the process returns to step S1, and the subsequent processes are repeated.

  On the other hand, when it is determined in step S5 that the face area image currently captured by the imaging device 11 is an image suitable for recognition processing, the determination unit 62 is supplied from the face area detection unit 22 in step S8. The face area image is supplied to the recognition processing unit 24 as a comparison image, and the process is terminated. As a result, the recognition processing unit 24 compares the optimal face area image from the determination unit 62 with the reference image stored in the reference image storage unit 25, thereby appearing in the face image captured by the imaging device 11. It is possible to perform an authentication process for authenticating whether or not the face (user) is the registered person.

  Next, a reference image registration process for registering a reference image serving as a reference in the face authentication process in the reference image storage unit 25 will be described with reference to the flowchart of FIG. This process is performed in advance before the comparative image acquisition process shown in FIG.

  First, in step S <b> 21, the communication unit 26 supplies a timing signal to the imaging device 11. And the face image imaged with the imaging device 11 with the timing signal from the communication part 26 is input into the frame memory 21, and the frame memory 21 memorize | stores the face image from the imaging device 11 in step S22. The process in step S21 is repeated several times as will be described later. In the first process, a face image captured by the imaging device 11 with automatic exposure is input to the frame memory 21. In the second and subsequent processes, the face image photographed at the aperture value F and the shutter speed T transmitted to the imaging device 11 in the process of step S27 described later is input to the frame memory 21.

  In step S 23, the face area detection unit 22 detects a face area from the face image supplied from the frame memory 21, and supplies face area information obtained as a result to the image determination unit 23.

  In step S <b> 24, the distribution creating unit 61 creates a luminance distribution of the face image supplied from the frame memory 21. In addition, the distribution creating unit 61 specifies a face region portion in the luminance distribution of the face image based on the face region information supplied from the face region detecting unit 22.

  In step S25, the determination unit 62 of the image determination unit 23 determines whether the luminance distribution of the face area image has been obtained a predetermined number of times by changing some exposure amounts (EV values) of the imaging device 11. In the image processing device 12, in the reference image registration process, the exposure amount is changed to, for example, a reference EV value, a + 2EV value, a + 1EV value, a -1EV value to a -3EV value, and a total of six face area image distributions Is stored in advance, and it is determined in step S25 whether or not the predetermined number of brightness distributions of the face area image have been obtained.

  If it is determined in step S25 that the luminance distribution of the face area image for the predetermined number of times has not been obtained, in step S26, the EV correction calculation unit 63 sets the shutter speed T so that the EV value has not been obtained yet. change.

  In step S <b> 27, the communication unit 26 transmits the aperture value F and the shutter speed T supplied from the EV correction calculation unit 63 of the image determination unit 23 to the imaging device 11. Thereafter, the processing returns to step S21, and the subsequent processing is repeated.

  On the other hand, when it is determined in step S25 that the luminance distribution of the face area image for a predetermined number of times has been obtained, in step S28, the determination unit 62 is obtained by repeatedly executing the processes in steps S21 to S24. An optimal face area image is selected as a reference image from the luminance distribution of face area images having different EV values, and is stored in the reference image storage unit 25, and the process ends. In step S <b> 28, the determination unit 62 out of the luminance distributions of the predetermined number of face area images obtained by the predetermined number of times of imaging, the central value 127 of the pixel values whose luminance distribution has a central value of 0 to 255 (256 gradations). And a face image having a large distribution spread (large variance) is selected as a reference image and stored in the reference image storage unit 25.

  FIG. 6 shows the luminance distribution of face area images with different EV values obtained by the reference image registration process of FIG.

  In the luminance distribution of the six face area images from -3EV value to + 2EV value shown in FIG. 6, when the image is captured with the -1EV value that is 1/2 the brightness than the automatic exposure (reference EV value). Are selected as a reference image and stored in the reference image storage unit 25.

  As described above, according to the image processing system 1 of FIG. 1, the reference image is obtained by using the luminance distribution of only the face area among the face images captured by the imaging device 11 including the face to be recognized. sign up. Further, by using the luminance distribution of only the face area, more specifically, the imaging device so that the error between the central value of the luminance distribution of the face area image and the central value of the luminance distribution of the reference image is within a predetermined range. Since the aperture value F or the shutter speed T of 11 is controlled, an image optimal for recognition processing can be obtained.

  Next, an embodiment in which the aperture value F and the shutter speed T are not adjusted by employing an image sensor having a wider dynamic range than the image sensor such as the above-described CMOS sensor or CCD sensor will be described. .

  FIG. 7 shows a second embodiment of the image processing system. In FIG. 7, portions corresponding to those in the first embodiment described above are denoted by the same reference numerals, and description thereof is omitted as appropriate.

  The image processing system 101 in FIG. 7 includes an HDRC (High Dynamic Range CMOS) (registered trademark) imaging device 111 having a very wide dynamic range of about 170 dB, and an image processing device 12 similar to the first embodiment. It is configured.

  As will be described later with reference to FIGS. 8 and 9, the HDRC image capturing apparatus 111 captures a face as a recognition target with a very wide dynamic range of about 170 dB using a logarithmic conversion type image sensor, and the image processing apparatus 112. Note that the HDRC imaging device 111 performs imaging based on the timing signal supplied from the image processing device 112, as in the imaging device 11 described above.

  The frame memory 21, the face area detection unit 22, the image determination unit 23, the recognition processing unit 24, the reference image storage unit 25, and the communication unit 26 of the image processing apparatus 12 in FIG. 7 are basically referred to FIG. The same processing as described is performed. However, in FIG. 7, the communication unit 26 supplies only the timing signal to the HDRC imaging device 111 and does not supply the aperture value F and shutter speed T to the HDRC imaging device 111.

  FIG. 8 shows a detailed configuration example of the HDRC imaging apparatus 111 of FIG.

  The HDRC imaging device 111 includes a lens 141, a diaphragm 142, a logarithmic conversion type image sensor 143, a logarithmic conversion unit 144, an A / D conversion unit 144, and an imaging control unit 45.

  Light from the subject that has passed through the lens 141 and the diaphragm 142 is incident on the logarithmic conversion type image sensor 143. The logarithmic conversion type image sensor 143 is configured by, for example, HDRC and captures images in accordance with a timing signal from the image capturing control unit 146. That is, the logarithmic conversion type image sensor 143 accumulates charges according to the brightness (illuminance) of incident light in synchronization with the timing signal, and supplies the accumulated charges to the logarithmic conversion unit 144.

  The logarithmic conversion unit 144 is configured by, for example, a plurality of MOSFETs (Metal Oxide Semiconductor Field Effect Transistors). The logarithmic conversion unit 144 uses the sub-threshold characteristic of the MOSFET to convert the charge supplied from the logarithmic conversion type image sensor 143 into the logarithm of the number of charges (the intensity of current) for each pixel (the amount of light of the subject). An analog electric signal converted into a voltage value approximately proportional to the logarithm) is generated. The logarithmic conversion unit 144 supplies the generated analog electrical signal to the A / D conversion unit 145.

The A / D conversion unit 145 converts the analog electrical signal supplied from the logarithmic conversion unit 144 into a digital signal having a predetermined number of bits and outputs the digital signal to the image processing device 12. Note that the A / D conversion unit 145 of the HDRC imaging device 111 employs a 14-bit number to cope with a wide dynamic range. Therefore, the pixel value of the face image takes a value in the range of 0 to the brightest 2 ¹⁴ −1.

  The imaging control unit 146 controls the timing of imaging by supplying the timing signal supplied from the image processing device 12 to the logarithmic conversion type image sensor 143.

  The HDRC imaging device 111 configured as described above outputs a digital image signal having a pixel value proportional to the logarithm of the brightness (incident light amount) of light incident on the logarithmic conversion type image sensor 143. Here, “proportional” includes those that are substantially proportional. The details of the logarithm conversion type image sensor are disclosed in, for example, Japanese Patent Publication No. 7-506932. In the logarithmic conversion type image sensor 143, it is possible to supply the logarithmic conversion unit 144 as it is without accumulating charges.

  FIG. 9 is a graph showing sensitivity characteristics of the logarithmic conversion type image sensor 143, the CCD sensor, the silver salt film, and the human eye. The horizontal axis in FIG. 9 indicates the logarithmic value of the illuminance (unit is lux) of the incident light, and the vertical axis indicates the sensitivity to the illuminance of the incident light. Curve L1 represents the sensitivity characteristic of the logarithmic conversion type image sensor 143, curve L2 represents the sensitivity characteristic of the CCD sensor, curve L3 represents the sensitivity characteristic of the silver salt film, and curve L4 represents the sensitivity characteristic of the human eye. Yes.

  As described above, the logarithmic conversion type image sensor 143 outputs an image signal having a pixel value proportional to the logarithm of the incident light quantity, thereby saturating the capacitances of the photodiodes, MOSFETs, and the like constituting the logarithmic conversion type image sensor 143. Without the image sensor, the subject can be imaged with a dynamic range of about 170 dB, which is wider than the human eye, from about 1 millilux to about 500 kilolux higher than the brightness of sunlight.

  Accordingly, the imaging device 111 using the logarithmic conversion type image sensor 143 does not generate luminance clipping in a luminance range that can be visually recognized by humans, and thus does not need to adjust the aperture value F or the shutter speed T to adjust the amount of incident light. . That is, the imaging device 111 can faithfully reproduce the detailed luminance distribution of the subject without adjusting the amount of incident light.

  For example, when a face is imaged in the daytime, even if the sun is within the angle of view, the image captured by the imaging device 111 can faithfully reproduce the sun and face luminance distribution without adjusting the amount of incident light. Reproduced image. Even when a face is imaged at night, the image captured by the imaging device 111 is an image that faithfully reproduces the luminance distribution of the area other than the face and the face without adjusting the amount of incident light.

  Note that the dynamic range of the logarithmic conversion type image sensor 143 is not limited to the above-described 170 dB, and a dynamic range corresponding to a required dynamic range such as about 100 dB or 200 dB can be used depending on the purpose of use.

  Further, in the CCD sensor and the silver salt film, as shown by the curves L2 and L3, the logarithmic conversion type image sensor 143 is compared with the case where the sensitivity characteristic is not proportional to the logarithm of the illuminance of incident light due to factors such as the gamma characteristic. Then, the sensitivity characteristic is proportional to the logarithm of the illuminance of incident light. From this, the imaging device 111 using the logarithmic conversion type image sensor 143 is not affected by the occurrence of luminance clipping, adjustment of the incident light amount, and gamma characteristics, so the pixel value of the image captured by the imaging device 111 is It also has a feature of reflecting the change in luminance of the subject and the movement of the subject almost faithfully.

  As described above, according to the image processing system 101 of FIG. 7, since the HDRC imaging apparatus 111 has a very wide dynamic range, it is necessary to adjust the aperture value F and the shutter speed T to adjust the amount of incident light. There is no.

However, as shown in FIG. 10, the face image output from the HDRC imaging device 111 is represented by 14 bits, that is, pixel values of 0 to 2 ¹⁴ −1. Only a few bits (2 or 3 bits in FIG. 10) of gradation are shown. That is, in the HDRC imaging device 111, the proportion of the pixel values in the face area is small due to the wide dynamic range.

  An A / D converter that performs A / D conversion processing is inexpensive because it is frequently used in a general electric circuit up to the number of bits of 8 bits, but one with a number of bits larger than 8 bits is expensive. In other words, the image processing system 101 using the HDRC imaging device 111 requires an expensive 14-bit A / D converter, but only a few bits are used for the face area that is actually required for the recognition process. There is no problem.

  FIG. 11 is a graph showing the result of examining the minimum number of gradations of pixel values required for performing accurate recognition processing.

  The horizontal axis of the graph shown in FIG. 11 represents the illuminance (unit is lux) of the subject, and the vertical axis on the right side of the graph represents the similarity between the reference image and the comparison image. From the experimental results, it is known that a highly reliable recognition process can be performed when the similarity is 550 or more indicated by a rough dotted line. Also, the vertical axis on the left side of the graph represents the pixel value of the face area image in 10 bits.

  In FIG. 11, a fine dotted line represents the maximum pixel value, and a two-dot chain line represents the minimum pixel value. A one-dot chain line represents a difference obtained by subtracting a minimum value from a maximum value of pixel values, and a solid line represents a similarity between the reference image and the comparison image.

  The graph shown in FIG. 11 indicates that the illuminance of the subject needs to be at least about 30 lux in order for the similarity to exceed 550. In the graph shown in FIG. 11, the difference between the pixel values when the similarity is 550 is about 175 as indicated by a one-dot chain line, and the reference image and the comparison image are 7 bits (128 bits). This indicates that a gradation greater than (gradation) is required.

Therefore, as a third embodiment, instead of the 14-bit A / D conversion unit, an 8-bit A / D conversion unit is employed, and as shown in FIG. Of the 2 ¹⁴ -1 pixel values, the number of gradations in the range from the minimum value to the maximum value of the pixel values in the face region is, for example, a predetermined ratio (for example, 80%) with respect to 8 bits (256 gradations). It is assumed that an electrical signal from the lower limit value M1 to the upper limit value M2 of the pixel value that is equal to or 90% is input to the 8-bit A / D conversion unit. In other words, the lower limit value M1 and the upper limit value M2 of the pixel value are set in a range from the minimum value to the maximum value of the pixel values in the face area among the pixel values of 0 to 2 ¹⁴ −1 obtained by the HDRC imaging apparatus. To K times (1 ≦ K). For example, when the predetermined ratio is 80%, the face area is represented by a pixel value of 256 × 0.8 = 204 gradations, and the above condition of about 175 or more is satisfied. it can.

  FIG. 13 shows a third embodiment of such an image processing system.

  An image processing system 201 in FIG. 13 includes an HDRC imaging device 211 and an image processing device 212. Note that, in FIG. 13, portions corresponding to those in the above-described embodiment are denoted by the same reference numerals, and description thereof is omitted as appropriate.

  In the image processing system 201, it is necessary to capture at least twice with the HDRC imaging device 211 when registering the reference image in the image processing device 212 and when inputting the comparison image into the image processing device 212. The first reference image or comparison image is used to determine a lower limit value M1 and an upper limit value M2 of pixel values that include a face area image at a predetermined ratio. The second reference image or comparison image is an image extracted according to the lower limit value M1 and the upper limit value M2 of the pixel value determined in the first face image, and is used as an image for authentication processing. Note that the second image captured and output by the HDRC imaging device 211 is also referred to as a corrected image. In the image processing device 212, whether the input image is the first image or the second image can be recognized by an instruction from the operation unit or the control unit.

  The HDRC imaging device 211 images a face as a recognition target with a very wide dynamic range of about 170 dB using a logarithmic conversion type imaging device, and converts an analog image signal obtained as a result into an 8-bit digital signal. To the image processing apparatus 212. Note that the HDRC imaging device 211 performs imaging based on the timing signal supplied from the image processing device 212, similarly to the imaging device 111 of FIG.

  The image processing device 212 includes a frame memory 21, a face area detection unit 22, a correction amount determination unit 221, a recognition processing unit 24, a reference image storage unit 25, and a communication unit 222.

  That is, the image processing device 212 is common to the image processing device 12 of FIG. 7 in that it includes the frame memory 21, the face area detection unit 22, the recognition processing unit 24, and the reference image storage unit 25. 7 is different from the image processing apparatus 12 of FIG. 7 in that a correction amount determination unit 221 and a communication unit 222 are provided instead of the image determination unit 23 and the communication unit 26.

  The correction amount determination unit 221 is supplied with the face image from the frame memory 21 and the face area information from the face area detection unit 22. The correction amount determination unit 221 specifies a face area image from the face image based on the face area information supplied from the face area detection unit 22. Then, the correction amount determination unit 221 calculates the minimum value and the maximum value of the pixel values of the face area image, and the pixels whose pixel value range of the calculated face area is, for example, 80% of the entire face area image. A lower limit value M1 and an upper limit value M2 are determined. Further, the correction amount determination unit 221 supplies the determined lower limit value M1 and upper limit value M2 of the pixel value to the communication unit 222.

  The communication unit 222 supplies the lower limit value M1 and the upper limit value M2 of the pixel value as the correction amount supplied from the correction amount determination unit 221 to the imaging device 211. The communication unit 26 also supplies a timing signal to the imaging unit 11 based on a control signal from an operation unit or a control unit (not shown).

  FIG. 14 shows a detailed configuration example of the HDRC imaging apparatus 211 of FIG.

  The HDRC imaging apparatus 211 includes a lens 141, a diaphragm 142, a logarithmic conversion type image sensor 143, a logarithmic conversion unit 144, a signal correction unit 241, an A / D conversion unit 242, and an imaging control unit 243.

  That is, the HDRC imaging device 211 in FIG. 14 is common to the HDRC imaging device 111 in FIG. 8 in that it has a lens 141, a diaphragm 142, a logarithmic conversion type image sensor 143, and a logarithmic conversion unit 144, and an A / D conversion unit. It differs from the HDRC imaging device 111 of FIG. 8 in having a signal correction unit 241, an A / D conversion unit 242, and an imaging control unit 243 instead of the 144 and the imaging control unit 45.

  The signal correction unit 241 is, for example, an analog signal processing circuit, and when the lower limit value M1 and the upper limit value M2 of the pixel value as the correction amount are supplied from the imaging control unit 243, the lower limit value M1 and the upper limit value of the pixel value M2 is converted to a voltage value before A / D conversion. Hereinafter, voltage values corresponding to the lower limit value M1 and the upper limit value M2 of the pixel value are referred to as a lower limit voltage value and an upper limit voltage value. The signal correction unit 241 extracts an analog signal from the lower limit voltage value to the upper limit voltage value from the analog electrical signal supplied from the logarithmic conversion unit 144 and supplies the analog signal to the A / D conversion unit 242. On the other hand, when the correction amount is not supplied from the imaging control unit 243, the signal correction unit 241 supplies the analog electric signal supplied from the logarithmic conversion unit 144 to the A / D conversion unit as it is.

  The A / D conversion unit 242 is, for example, an A / D converter, converts an analog electrical signal supplied from the logarithmic conversion unit 144 into an 8-bit digital signal, and outputs the converted signal to the image processing apparatus 212 as an image signal. .

  The imaging control unit 243 controls the timing of imaging by supplying the timing signal supplied from the image processing device 212 to the logarithmic conversion type image sensor 143. Further, when the lower limit value M1 and the upper limit value M2 of the pixel value as the correction amount are supplied from the image processing device 212, the imaging control unit 243 supplies them to the signal correction unit 241.

  FIG. 15 shows a detailed configuration example of the correction amount determination unit 221 of FIG.

  The correction amount determination unit 221 includes a distribution creation unit 261 and an upper / lower limit determination unit 262.

  The distribution creating unit 261 creates a luminance distribution of the face image supplied from the frame memory 21. The upper and lower limit determination unit 262 specifies a face region portion in the luminance distribution of the face image based on the face region information supplied from the face region detection unit 22. Then, the upper and lower limit determination unit 262 determines that the range of the pixel value of the face area portion is a predetermined ratio (for example, 80%) from the minimum value and the maximum value of the pixel value of the specified face area portion. A lower limit value M1 and an upper limit value M2 are determined. The determined lower limit value M1 and upper limit value M2 of the pixel value are supplied to the communication unit 222.

  The recognized image acquisition process of the image processing system 201 will be described with reference to the flowchart of FIG.

  First, in step S61, a timing signal is supplied from the image processing device 212 to the HDRC imaging device 211, and the logarithmic conversion type image sensor 143 of the HDRC imaging device 211 images a face as a recognition target. Then, an analog electric signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the signal correction unit 241.

  In step S <b> 62, the signal correction unit 241 supplies the analog signal from the logarithmic conversion unit 144 to the A / D conversion unit 242 as it is. In step S62, the A / D conversion unit 242 converts the analog electrical signal from the signal correction unit 241 into an 8-bit digital signal, and outputs the converted signal to the image processing apparatus 212 as an image signal.

  In step S <b> 63, the frame memory 21 of the image processing device 212 stores the face image from the HDRC imaging device 211. The face image stored in the frame memory 21 is supplied to the face area detection unit 22 and the correction amount determination unit 221.

  In step S 64, the face area detection unit 22 detects a face area from the face image supplied from the frame memory 21, and supplies the face area information obtained as a result to the correction amount determination unit 221.

  In step S <b> 65, the distribution creation unit 261 of the correction amount determination unit 221 creates the brightness distribution of the face image supplied from the frame memory 21.

  In step S <b> 66, the upper and lower limit value determination unit 262 of the correction amount determination unit 221 specifies the face region portion in the luminance distribution of the face image based on the face region information supplied from the face region detection unit 22. Then, the upper and lower limit value determining unit 262 determines that the range of the pixel values of the face area portion is a predetermined ratio (for example, 80%) of the whole from the minimum value and the maximum value of the pixel values of the specified face area portion. The lower limit value M1 and the upper limit value M2 of the pixel value are determined.

  In step S <b> 67, the lower limit value M <b> 1 and the upper limit value M <b> 2 of the determined pixel value are supplied to the communication unit 222, and the communication unit 222 supplies them to the HDRC imaging device 211.

  In step S <b> 68, the imaging control unit 243 of the HDRC imaging device 211 supplies the upper limit value M <b> 1 and the lower limit value M <b> 2 of the pixel value supplied from the image processing device 212 to the signal correction unit 241. Then, the signal correction unit 241 converts the lower limit value M1 and the upper limit value M2 of the pixel value into voltage values (lower limit voltage value and upper limit voltage value) before A / D conversion.

  In step S69, a timing signal is supplied from the image processing device 212 to the HDRC imaging device 211, and the logarithmic conversion type image sensor 143 of the HDRC imaging device 211 images a face. Then, an analog electric signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the signal correction unit 241.

  In step S70, the signal correction unit 241 of the HDRC imaging apparatus 211 extracts an analog signal from the lower limit voltage value to the upper limit voltage value from the analog electric signal supplied from the logarithmic conversion unit 144, and the A / D conversion unit. 242 is supplied.

  In step S71, the A / D conversion unit 242 of the HDRC imaging apparatus 211 converts the analog electrical signal from the signal correction unit 241 into an 8-bit digital signal, and outputs the converted signal to the image processing device 212 as an image signal.

  In step S <b> 72, the frame memory 21 of the image processing device 212 stores the face image from the HDRC imaging device 211. The face image stored in the frame memory 21 is supplied to the correction amount determination unit 221.

  In step S73, the correction amount determination unit 221 supplies the face image supplied from the frame memory 21 to the recognition processing unit 24, and ends the process. Accordingly, the recognition processing unit 24 can perform authentication processing using the face area image from the correction amount determination unit 221 and the reference image stored in the reference image storage unit 25.

  As described above, according to the recognition image acquisition process of the image processing system 201, the range of the maximum value is obtained from the minimum value of the pixel values of the face area portion from the face image obtained by the first imaging. In this imaging, only the image signal (analog electrical signal) in that range is input to the A / D conversion unit 242, so that the number of gradations is sufficient to perform recognition processing of about 200 gradations. A face area image can be obtained. That is, an image optimal for recognition processing can be obtained.

  The HDRC image pickup apparatus 211 used for image pickup in the image processing system 201 can pick up an object with a wide dynamic range of about 170 dB. Therefore, it is necessary to adjust the aperture value F and the shutter speed T to adjust the amount of incident light. There is no. Further, since the A / D converter 242 of the HDRC image pickup apparatus 211 only needs to convert it into an 8-bit digital signal, it is not necessary to use an expensive A / D converter having a larger number of bits than 8 bits, and the cost is reduced. Can be suppressed.

  In the image processing system 201, the reference image can be similarly obtained by the recognized image acquisition process of FIG. 16 and stored in the reference image storage unit 25.

  In the image processing system 201 described above, the correction amount is determined by the image processing device 212 and supplied to the HDRC image pickup device 211. However, the correction amount is obtained by the HDRC image pickup device itself and is represented by 8-bit gradation. It is also possible to output a corrected image.

  FIG. 17 shows a configuration example of an HDRC imaging apparatus in which the HDRC imaging apparatus itself obtains a correction amount and outputs a corrected image represented by 8-bit gradation (fourth embodiment). ). In FIG. 17, parts corresponding to those in the above-described embodiment are given the same reference numerals, and the description thereof is omitted as appropriate.

  The HDRC imaging device 301 of FIG. 17 includes the lens 141, the diaphragm 142, the logarithmic conversion type image sensor 143, the logarithmic conversion unit 144, the signal correction unit 241, and the A / D conversion unit 242, and therefore the HDRC imaging device of FIG. The HDRC image pickup apparatus 211 is different from the HDRC image pickup apparatus 211 of FIG. However, the frame memory 21, the face area detection unit 22, and the correction amount determination unit 221 of the HDRC imaging apparatus 301 are the same as those included in the image processing apparatus 212 of FIG.

  The switch 311 switches whether the electrical signal from the A / D conversion unit 242 is output to a subsequent device (not shown) or supplied to the frame memory 21 under the control of the imaging control unit 312. When the electrical signal from the A / D conversion unit 242 is output to a subsequent device, the terminal unit 311a is selected, and when the electrical signal is supplied to the frame memory 21, the terminal unit 311b is selected.

  The imaging control unit 312 controls the imaging timing by supplying a timing signal to the logarithmic conversion type image sensor 143. In addition, the imaging control unit 312 causes the switch 311 to switch between outputting an electrical signal from the A / D conversion unit 242 to a subsequent device (not shown) or supplying the electrical signal to the frame memory 21. Note that the imaging control unit 312 may periodically supply a timing signal or switch the switch 311, or supply a timing signal or switch the switch 311 based on a control signal from another device. You may let them. In FIG. 17, a control line from the imaging control unit 312 to the switch 311 is not shown.

  FIG. 18 is a flowchart of the recognized image output process of the HDRC imaging apparatus 301 in FIG. 17 for outputting a corrected image as a recognized image.

  In the recognition image output process, first, in step S101, the imaging control unit 312 sets the switch 311 to the frame memory 21 side, that is, the terminal unit 311b.

  In step S <b> 102, the logarithmic conversion type image sensor 143 images a face as a recognition target based on the timing signal from the imaging control unit 312. Then, an analog electric signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the signal correction unit 241.

  In step S <b> 103, the signal correction unit 241 supplies the analog signal from the logarithmic conversion unit 144 to the A / D conversion unit 242 as it is. Further, in step S 103, the A / D conversion unit 242 converts the analog electric signal into an 8-bit digital signal and supplies it to the switch 311. In the switch 311, since the terminal unit 311 b is selected, the electrical signal supplied from the A / D conversion unit 242 is supplied to the frame memory 21.

  In step S <b> 104, the frame memory 21 stores the face image supplied from the A / D conversion unit 242 via the switch 311.

  In step S <b> 105, the face area detection unit 22 detects a face area from the face image supplied from the frame memory 21, and supplies face area information obtained as a result to the correction amount determination unit 221.

  In step S <b> 106, the distribution creation unit 261 of the correction amount determination unit 221 creates the luminance distribution of the face image supplied from the frame memory 21.

  In step S <b> 107, the upper and lower limit value determination unit 262 of the correction amount determination unit 221 specifies a face area portion in the luminance distribution of the face image based on the face area information supplied from the face area detection unit 22. Then, the upper and lower limit determination unit 262 determines that the range of the pixel value of the face area portion is a predetermined ratio (for example, 80%) from the minimum value and the maximum value of the pixel value of the specified face area portion. The lower limit value M1 and the upper limit value M2 of the correct pixel value are determined and supplied to the signal correction unit 241.

  In step S108, the signal correction unit 241 converts the lower limit value M1 and the upper limit value M2 of the pixel value into voltage values (lower limit voltage value and upper limit voltage value) before A / D conversion.

  In step S109, the imaging control unit 312 sets the switch 311 to the external output side, that is, the terminal unit 311a.

  In step S <b> 110, the logarithmic conversion type image sensor 143 images the face based on the timing signal from the imaging control unit 312. Then, an analog electric signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the signal correction unit 241.

  In step S <b> 111, the signal correction unit 241 extracts an analog signal from the lower limit voltage value to the upper limit voltage value from the analog electric signal supplied from the logarithmic conversion unit 144 and supplies the analog signal to the A / D conversion unit 242.

  In step S112, the A / D conversion unit 242 converts the analog electrical signal from the signal correction unit 241 into an 8-bit digital signal, outputs it as an image signal, and ends the processing.

  As described above, the range of the maximum value is obtained from the minimum value of the pixel values of the face region from the face image obtained by the first imaging, and the upper limit value M2 from the lower limit value M1 at which the range becomes a predetermined ratio. Only the image signal (analog electrical signal) is input to the A / D conversion unit 242 in the second imaging, so that the range of pixel values in the face region has a sufficient number of gradations for recognition processing. It is possible to output a face area image having the same. That is, an image optimal for recognition processing can be obtained.

  In the HDRC imaging apparatus 301 of FIG. 17, it is necessary to perform imaging twice, but FIG. 19 shows an example of the configuration of the HDRC imaging apparatus that can be performed by one imaging (fifth embodiment). Form). In FIG. 19, the same reference numerals are given to the portions corresponding to the above-described embodiment, and the description thereof will be omitted as appropriate.

  19 includes a frame memory 21, a face area detection unit 22, a lens 141, a diaphragm 142, a logarithmic conversion type image sensor 143, a logarithmic conversion unit 144, a correction amount determination unit 221, a signal correction unit 241, 17 and in common with the HDRC imaging apparatus 301 in FIG. 17, a delay unit 411, a logarithmic conversion unit 144, a frame memory 21, and a logarithmic conversion unit 144 and a signal correction unit 241 are provided. 17 is different from the HDRC imaging apparatus 301 in FIG. 17 in that an A / D conversion unit 412 is provided between the two.

  The delay unit 411 delays the analog electrical signal supplied from the logarithmic conversion unit 144 for a predetermined time, and then supplies it to the signal correction unit 241. Specifically, the delay unit 411 receives the analog value supplied from the logarithmic conversion unit 144 until the correction amount determination unit 221 calculates the lower limit value M1 and the upper limit value M2 of the pixel value and supplies them to the signal correction unit 241. Delay the electrical signal.

  The A / D conversion unit 412 converts the analog electric signal supplied from the logarithmic conversion unit 144 into an 8-bit digital signal and supplies the converted signal to the frame memory 21.

  Next, a recognition image output process by the HDRC imaging apparatus 401 of FIG. 19 will be described with reference to the flowchart of FIG.

  First, in step S <b> 161, the logarithmic conversion type image sensor 143 images a face as a recognition target based on a timing signal from the imaging control unit 243. Then, an analog electrical signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the delay unit 411 and the A / D conversion unit 412.

  In step S <b> 162, the delay unit 411 delays the analog electrical signal supplied from the logarithmic conversion unit 144 for a predetermined time, and then supplies the analog electrical signal to the signal correction unit 241.

  In step S <b> 163, the A / D conversion unit 412 converts the analog electrical signal into an 8-bit digital signal and supplies it to the frame memory 21.

  In step S164, the frame memory 21 stores the face image supplied from the A / D conversion unit 412.

  In step S 165, the face area detection unit 22 detects a face area from the face image supplied from the frame memory 21, and supplies face area information obtained as a result to the correction amount determination unit 221.

  In step S166, the distribution creation unit 261 of the correction amount determination unit 221 creates the luminance distribution of the face image supplied from the frame memory 21.

  In step S 167, the upper and lower limit value determination unit 262 of the correction amount determination unit 221 specifies a face area portion in the luminance distribution of the face image based on the face area information supplied from the face area detection unit 22. Then, the upper and lower limit determination unit 262 determines that the range of the pixel value of the face area portion is a predetermined ratio (for example, 80%) from the minimum value and the maximum value of the pixel value of the specified face area portion. The lower limit value M1 and the upper limit value M2 of the correct pixel value are determined and supplied to the signal correction unit 241.

  In step S168, the signal correction unit 241 converts the lower limit value and the upper limit value of the pixel value into voltage values (lower limit voltage value and upper limit voltage value) before A / D conversion.

  In step S <b> 169, the signal correction unit 241 extracts an analog signal from the lower limit voltage value to the upper limit voltage value from the analog electric signal supplied from the delay unit 411, and supplies the analog signal to the A / D conversion unit 242.

  In step S170, the A / D conversion unit 242 converts the analog electrical signal from the signal correction unit 241 into an 8-bit digital signal, outputs it as an image signal, and ends the processing.

  As described above, the range of the maximum value is obtained from the minimum value of the pixel values of the face region portion, and the signal in the range is extracted from the delayed analog electric signal and input to the A / D converter 242. By doing so, it is possible to output a face area image having a sufficient number of gradations for performing the recognition process. That is, an image optimal for recognition processing can be obtained.

  In the first to fifth embodiments described above, an example has been described in which a face area as a recognition target is detected, and an image in which the luminance distribution of the face area image is optimal for recognition processing is acquired. However, when it is difficult to detect the face area itself, the user may designate the face area with a cursor or the like. Therefore, the image processing system designates a predetermined area in the image obtained by imaging with the HDRC imaging device, and an image in which the luminance distribution of the designated predetermined area is optimal for recognition processing. Can be requested.

  FIG. 21 shows a configuration example of an image processing system that obtains an image in which the luminance distribution of a specified predetermined region is optimal for recognition processing (sixth embodiment). In FIG. 21, portions corresponding to those in FIG. 13 are denoted by the same reference numerals, and description thereof is omitted as appropriate.

  An image processing system 501 in FIG. 21 includes an HDRC imaging device 211 and an image processing device 511. The image processing apparatus 511 includes a frame memory 21, a recognition processing unit 24, a reference image storage unit 25, a correction amount determination unit 221, a communication unit 222, an area designation unit 521, and a designated area detection unit 522.

  Compared with the image processing system 201 in FIG. 13, the image processing system 501 in FIG. 21 includes the HDRC imaging device 211, the frame memory 21, the recognition processing unit 24, the reference image storage unit 25, the correction amount determination unit 221, and the communication. The area specifying unit 521 and the specified area detecting unit 522 are different. In the sixth embodiment, an image captured by the HDRC image capturing device 211 and output to the image processing device 511 is referred to as a captured image in distinction from a face image.

  The area designating unit 521 designates an area to be subjected to recognition processing from the image supplied from the HDRC imaging device 211. For example, the area designation unit 521 displays an area designation frame on the display (not shown) together with the captured image supplied from the HDRC imaging apparatus 211, and allows the user to set the area designation frame to a desired position and size.

  The designated area detection unit 522 stores the area designated by the area designation unit 521 as a target area for recognition processing. Also, the designated area detection unit 522 detects the area designated by the area designation unit 521 from the captured image supplied from the frame memory 21, and supplies the designated area information obtained as a result to the correction amount determination unit 221.

  Next, the area designation process by the image processing apparatus 511 will be described with reference to the flowchart of FIG.

  In step S181, the area specifying unit 521 displays an area specifying frame for specifying an area on the display.

  In step S182, the area designating unit 521 determines whether or not an area has been determined, and waits until it is determined that the area has been determined. For example, the user sets the area designation frame displayed on the display to a desired position and size, and operates the confirmation button or the like when the decision is made. As a result, in step S182, it is determined that the region has been determined.

  If it is determined in step S182 that the area has been determined, in step S183, the designated area detection unit 522 stores the area designated by the area designation unit 521 as a target area for recognition processing.

  Next, the recognized image acquisition process of the image processing system 501 will be described with reference to the flowchart of FIG.

  First, in step S201, a timing signal is supplied from the image processing device 511 to the HDRC imaging device 211, and the logarithmic conversion type image sensor 143 of the HDRC imaging device 211 images the subject. Then, an analog electric signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the signal correction unit 241.

  In step S202, the signal correction unit 241 supplies the analog signal from the logarithmic conversion unit 144 to the A / D conversion unit 242 as it is. Further, in step S202, the A / D conversion unit 242 converts the analog electric signal into an 8-bit digital signal and outputs it as an image signal to the image processing device 511.

  In step S <b> 203, the frame memory 21 of the image processing apparatus 511 stores the captured image from the HDRC imaging apparatus 211. The captured image stored in the frame memory 21 is supplied to the designated area detection unit 522 and the correction amount determination unit 221.

  In step S <b> 204, the designated area detection unit 522 detects the designated area designated by the above-described area designation processing from the captured image supplied from the frame memory 21, and designates the designated area information obtained as a result as the correction amount determination unit 221. To supply.

  In step S <b> 205, the distribution creation unit 261 of the correction amount determination unit 221 creates a luminance distribution of the captured image supplied from the frame memory 21.

  In step S <b> 206, the upper / lower limit determination unit 262 of the correction amount determination unit 221 specifies a designated region portion in the luminance distribution of the captured image based on the designated region information supplied from the designated region detection unit 22. Then, the upper and lower limit determination unit 262 determines the pixels in which the range of the pixel value in the designated area portion is a predetermined ratio (for example, 80%) from the minimum value and the maximum value of the pixel value in the specified designated area portion. A lower limit value M1 and an upper limit value M2 are determined.

  In step S <b> 207, the lower limit value M <b> 1 and the upper limit value M <b> 2 of the determined pixel value are supplied to the communication unit 222, and the communication unit 222 supplies them to the HDRC imaging device 211.

  In step S <b> 208, the imaging control unit 243 of the HDRC imaging device 211 supplies the upper limit value M <b> 1 and the lower limit value M <b> 2 of the pixel value supplied from the image processing device 511 to the signal correction unit 241. Then, the signal correction unit 241 converts the lower limit value and the upper limit value of the pixel value into voltage values (lower limit voltage value and upper limit voltage value) before A / D conversion.

  In step S209, a timing signal is supplied from the image processing device 511 to the HDRC imaging device 211, and the logarithmic conversion type image sensor 143 of the HDRC imaging device 211 images the subject. Then, an analog electric signal obtained by imaging is supplied from the logarithmic conversion unit 144 to the signal correction unit 241.

  In step S210, the signal correction unit 241 of the HDRC imaging apparatus 211 extracts an analog signal from the lower limit voltage value to the upper limit voltage value from the analog electric signal supplied from the logarithmic conversion unit 144, and the A / D conversion unit 242 is supplied.

  In step S211, the A / D conversion unit 242 of the HDRC imaging apparatus 211 converts the analog electrical signal from the signal correction unit 241 into an 8-bit digital signal and outputs the converted signal to the image processing apparatus 511 as an image signal.

  In step S <b> 212, the frame memory 21 of the image processing apparatus 511 stores the captured image from the HDRC imaging apparatus 211. The captured image stored in the frame memory 21 is supplied to the correction amount determination unit 221.

  In step S213, the correction amount determination unit 221 supplies the captured image supplied from the frame memory 21 to the recognition processing unit 24, and ends the process.

  As described above, according to the recognized image acquisition processing of the image processing system 501, the range of the maximum value is obtained from the minimum value of the pixel value of the designated area portion designated in advance from the photographed image obtained by the first imaging. In the second imaging, only the image signal (analog electrical signal) within that range is input to the A / D converter 242. Therefore, the designated area image having a sufficient number of gradations for performing the recognition process. Can be obtained. That is, an image optimal for recognition processing can be obtained.

  Further, since the HDRC imaging apparatus 211 used for imaging in the image processing system 501 can capture a subject with a wide dynamic range of about 170 dB, it is necessary to adjust the aperture value F and the shutter speed T to adjust the amount of incident light. There is no. Since it is not necessary to use an expensive A / D converter having more bits than 8 bits for the A / D converter 242 of the HDRC imaging apparatus 211, the cost can be suppressed.

  The HDRC imaging apparatus 601 in FIG. 24 and the HDRC imaging apparatus 701 in FIG. 25 detect the designated area instead of the face area with the HDRC imaging apparatus 301 in FIG. 17 and the HDRC imaging apparatus 401 in FIG. The example of a structure in the case of having been shown is shown.

  In the HDRC imaging device 601 in FIG. 24 and the HDRC imaging device 701 in FIG. 25, instead of the face region detection unit 22 in the HDRC imaging device 301 in FIG. 17 and the HDRC imaging device 401 in FIG. An area detection unit 522 is provided. The area designation unit 521 and the designated area detection unit 522 in FIGS. 24 and 25 are the same as those in FIG. Further, in the recognized image output processing of the HDRC imaging device 601 in FIG. 24 and the HDRC imaging device 701 in FIG. 25, the processing in step S105 in FIG. 18 and the processing in step S165 in FIG. Therefore, the description is omitted. The area designating process for designating an area is the same as the process described with reference to FIG.

  Also in the HDRC imaging device 601 in FIG. 24 and the HDRC imaging device 701 in FIG. 25, the range of the maximum value is obtained from the minimum value of the pixel value in the designated area portion, and the lower limit value M1 to the upper limit value M2 at which the range becomes a predetermined ratio. By extracting only this signal from the delayed analog electrical signal and inputting it to the A / D converter 242, it is possible to output a designated area image having a sufficient number of gradations for performing recognition processing. it can. That is, an image optimal for recognition processing can be obtained.

  The series of processes described above can be executed by hardware or can be executed by software. When a series of processing is executed by software, a program constituting the software may execute various functions by installing a computer incorporated in dedicated hardware or various programs. For example, it is installed from a network or a recording medium into a general-purpose personal computer or the like.

  FIG. 26 is a block diagram illustrating a configuration example of a computer that executes the above-described series of processing by a program. A CPU (Central Processing Unit) 801 executes various processes according to a program stored in a ROM (Read Only Memory) 802 or a storage unit 808. A RAM (Random Access Memory) 803 appropriately stores programs executed by the CPU 801, data, and the like. These CPU 801, ROM 802, and RAM 803 are connected to each other via a bus 804.

  An input / output interface 805 is also connected to the CPU 801 via the bus 804. Connected to the input / output interface 805 are an input unit 806 made up of a keyboard, mouse, microphone, etc., a display made up of a CRT (Cathode Ray Tube), LCD (Liquid Crystal display), etc., and an output unit 807 made up of speakers. The CPU 801 executes various processes in response to a command input from the input unit 806. Then, the CPU 801 outputs the processing result to the output unit 807.

  A storage unit 808 connected to the input / output interface 805 includes, for example, a hard disk, and stores programs executed by the CPU 801 and various data. A communication unit 809 communicates with an external device connected directly or via a network such as the Internet or a local area network.

  Note that the communication unit 809 may perform wireless or wired communication, or may be capable of both wireless and wired communication. Further, the communication method is not particularly limited. For example, in the case of wireless, wireless LAN (Local Area Network) such as IEEE (The Institute of Electrical and Electronic Engineers) 802.11a, 802.11b, and 802.11g, or various types such as Bluetooth Wireless communication systems can be used. Similarly, in the case of a wired connection, a wired communication system conforming to IEEE1394, Ethernet (registered trademark), USB (Universal Serial Bus), or the like can be used.

  The drive 810 connected to the input / output interface 805 drives a removable medium 821 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and drives the program or data recorded therein. Get etc. The acquired program and data are transferred to and stored in the storage unit 808 as necessary. Further, the program and data may be acquired via the communication unit 809 and stored in the storage unit 808.

  The imaging unit 811 includes an imaging device such as a CCD, CMOS, or HDRC, images a subject, and supplies image data of the captured subject to the CPU 801 or the like via the input / output interface 805.

  In the present specification, the steps described in the flowcharts are executed in parallel or individually even if they are not necessarily processed in time series, as well as processes performed in time series in the described order. It also includes processing.

  Further, in this specification, the system represents the entire apparatus constituted by a plurality of apparatuses.

  The embodiment of the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the gist of the present invention.

It is a figure which shows the structural example of one Embodiment of the image processing system to which this invention is applied. It is a block diagram which shows the detailed structural example of the imaging device of FIG. It is a block diagram which shows the detailed structural example of the image determination part of an image processing apparatus. It is a flowchart explaining the comparative image acquisition process of an image processing apparatus. It is a flowchart explaining the reference | standard image registration process of an image processing apparatus. It is a figure which shows the luminance distribution of a face image. It is a figure which shows the structural example of other embodiment of an image processing system. It is a figure which shows the detailed structural example of the HDRC imaging device of FIG. It is a figure which shows the sensitivity characteristic of a logarithm conversion type image sensor, a CCD sensor, a silver salt film, and a human eye. It is a figure explaining the luminance distribution of the face image output from a HDRC imaging device. It is a figure explaining the gradation number which performs a recognition process. It is a figure explaining the gradation number which performs a recognition process. It is a figure which shows the structural example of other embodiment of an image processing system. It is a figure which shows the detailed structural example of the HDRC imaging device of FIG. It is a block diagram which shows the detailed structural example of the correction amount determination part of FIG. It is a flowchart explaining the recognition image acquisition process of the image processing system of FIG. It is a figure which shows the structural example of embodiment of an HDRC imaging device. It is a flowchart explaining the recognition image output process of the HDRC imaging device of FIG. It is a figure which shows the structural example of other embodiment of an HDRC imaging device. FIG. 20 is a flowchart for describing recognition image output processing of the HDRC imaging apparatus of FIG. 19. It is a figure which shows the structural example of other embodiment of an image processing system. It is a flowchart explaining the area | region designation | designated process of the image processing system of FIG. It is a flowchart explaining the recognition image acquisition process of the image processing system of FIG. It is a figure which shows the structural example of other embodiment of an HDRC imaging device. It is a figure which shows the structural example of other embodiment of an HDRC imaging device. It is a block diagram which shows the structural example of one Embodiment of the computer to which this invention is applied.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image processing system 11 Imaging device 22 Face area detection part 23 Image determination part 26 Communication part 61 Distribution creation part 62 Determination part 63 EV correction calculation part 101 Image processing system 111 HDRC imaging device 201 Image processing system 211 HDRC imaging device 221 Correction amount Determination unit 222 Communication unit 241 Signal correction unit 242 A / D conversion unit 261 Distribution creation unit 262 Upper / lower limit determination unit 301, 401 HDRC imaging device 501 Image processing system 521 Region designation unit 522 Designated region detection unit 601, 701 HDRC imaging device

Claims (7)

Reference image storage means for storing a reference image serving as a reference for authenticating whether or not the face included in the image captured by the imaging device is a registered person;
Face area detecting means for detecting a face area from the image;
A histogram creating means for creating a histogram of pixel values of each pixel in the face area in the image;
It is determined whether or not an error between a pixel value histogram of each pixel in the face area of the image and a pixel value histogram of each pixel of the reference image is within a predetermined range, and the error is not within the predetermined range. to be within a predetermined when a determining means for determining an exposure amount of the imaging device imaging the image,
Control means for controlling the imaging device so as to achieve the determined exposure amount ;
By comparing the image in which the error is determined to be within the predetermined range and the reference image stored in the reference image storage unit, the face included in the image captured by the imaging device is An image processing apparatus comprising authentication means for authenticating whether or not the person is registered .   Reference image registration means for storing the reference image in the reference image storage means,
  When the reference image registration means stores the reference image in the reference image storage means, the control means performs a plurality of controls so that each has a different exposure amount,
  The reference image registration unit acquires a plurality of images captured by the imaging device with different exposure amounts, and uses the acquired one of the plurality of images as the reference image in the reference image storage unit. Remember
  The image processing apparatus according to claim 1.   The reference image registration unit is configured such that, in the acquired plurality of images, a center value of a histogram of pixel values of each pixel of the face area of each image is close to a center value of a range that the pixel value can take. An image having a large distribution spread is stored in the reference image storage means as the reference image.
  The image processing apparatus according to claim 2. It said determination means, by determining the shutter speed or F value of the imaging device, an image processing apparatus according to any one of claims 1 to 3 for determining the exposure amount of the imaging device. The determination unit is configured to determine whether an error between a center value of a histogram of pixel values of each pixel of the face area and a center value of a histogram of pixel values of each pixel of the reference image is within a predetermined range. the image processing apparatus according to any one of claims 1 to determine the amount of exposure to 3. A reference image that serves as a reference for authenticating whether or not the face included in the image captured by the imaging device is a registered person is stored in a predetermined storage unit,
Detecting a face region from the image,
Creating a histogram of pixel values of each pixel in the face region in the image;
It is determined whether or not an error between a pixel value histogram of each pixel in the face area of the image and a pixel value histogram of each pixel of the reference image is within a predetermined range, and the error is not within the predetermined range. to be within a predetermined when determines the exposure amount of the imaging device imaging the image,
Controlling the imaging device to achieve the determined exposure amount ;
The face included in the image captured by the imaging device is registered by comparing the image in which the error is determined to be within the predetermined range and the reference image stored in the storage unit. An image processing method including the step of authenticating whether or not the person is a person . A reference image that serves as a reference for authenticating whether or not the face included in the image captured by the imaging device is a registered person is stored in a predetermined storage unit,
Detecting a face region from the image,
Creating a histogram of pixel values of each pixel in the face region in the image;
It is determined whether or not an error between a pixel value histogram of each pixel in the face area of the image and a pixel value histogram of each pixel of the reference image is within a predetermined range, and the error is not within the predetermined range. to be within a predetermined when determines the exposure amount of the imaging device imaging the image,
Controlling the imaging device to achieve the determined exposure amount ;
The face included in the image captured by the imaging device is registered by comparing the image in which the error is determined to be within the predetermined range and the reference image stored in advance in the storage unit. A program that causes a computer to execute a process including a step of authenticating whether or not the user is a registered person .

Images