JPH066747A - Id photo system - Google Patents

Abstract

PURPOSE:To reduce data volume required for a background part and to record and store an ID face image without increasing memory capacity by substituting representative compressed data previously formed based upon the partial data of a background image area for the background image area. CONSTITUTION:Face image data in photographing are inputted to a frame memory 25, and at a moment of requiring the freezing of the face image, a signal is applied from a controller 30 to the memory 25 through a keyboard 31 to successively stop writing and freeze the face image. In the case of printing out an ID card image, data are previously stored in an external memory 43 and transferred to an image processing part 33. A memory address control part 32 is controlled by a controller 30 and data are written on a fixed position in the memory 25. Then, the face image of the compressed image is read out from the memory 43, extended and written in a prescribed position of the memory 25 from the processing part 33. After forming an ID card image on the memory 25, the image is sent to a video printer 29 by the controller and printed out.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an ID photo system for picking up an ID face image.

[0002]

2. Description of the Related Art Conventionally, the use of image information has been limited because it requires a large amount of information as compared with character and numerical information, but recently, the progress and development of equipment for handling image information has advanced. As a result, the restrictions are decreasing.

The ID field is known as one of the fields in which images play an important role. And such ID
In the field, the importance of facial photographs has long been recognized, and therefore, silver lead photographs, instant photographs, etc. have been used for ID purposes as means for obtaining facial photographs.

However, since such silver lead photographs and instant photographs consist of information on photographic paper,
There is information only in the print, and it is difficult to take out this information, file it, process it, and copy it, and its application is extremely limited. Therefore, in order to eliminate such a problem, it has been considered to convert facial photograph information into an electric signal and handle it. FIG. 10 is based on such an idea and shows a diagram quoted from Japanese Patent Laid-Open No. 63-316275.

FIG. 1 shows an ID card producing apparatus. When a subject 1 is imaged by a color TV camera 2, this image pickup signal is converted into a digital multi-gradation video image signal by an A / D conversion circuit 3. It is temporarily recorded in the frame memory 4. Then, with respect to the video image signal recorded in the frame memory 4, an operation for removing color turbidity due to an incorrect spectral absorption element of the color material applied to the full-color video printer by the image signal processing circuit 5 under the control of the controller 6 and the like. After that, a graphics image or a character image is generated, a person image of the subject 1 and predetermined information are combined and edited to generate an image of an ID card, which is printed out through the full-color video printer unit 7. ing.

[0006] On the other hand, as an example of using a face photograph, there is a database for managing the information of members who belong to an organization such as a company or a school. An example of such a database system is known as shown in FIG. In this device, the video imaging devices 8, 9,
The video image signal from 10 is taken into the image processor 11, which is displayed on the monitor 12 and stored in the data file 13 so that it can be printed out by the video printer 14 if necessary.

By the way, in the case of a facial photograph, the video image handled in this way requires a storage capacity of several hundred kilobytes per person. For this reason, when recording and storing a face picture as a video image, the storage capacity of a generally used storage medium such as a magneto-optical recording disk or a hard disk will be filled with a video image for about 1,000 people.

For this reason, as a practical problem, it is urgently necessary to record and save as many face photographs as possible, and as a means for solving this problem, an image of an encoding system using a high-efficiency encoding technique is used. Data compression is being considered.

[0009]

However, the image data compression by such an encoding method is also performed by JPEG (J
Oint Picture Expert Grou
p) or a data compression method that deviates from the standardized method, the compatibility of the image data cannot be maintained, so that the data can not be shared and cannot be used in other systems, and the future image data can be used. There was a risk of leaving a problem.

Further, in the case of a face photograph, the background portion behind the face image is often a bright and dark image due to uneven lighting, and noise is present. Therefore, the background portion is the same as the face image portion. However, there is also a problem in that it is difficult to increase the data compression rate when the image data is compressed.

Further, in the conventional encoding method, since it corresponds to a general image having various information in each part of one image as an image object, it is complicated when applied to a face image for ID. However, there were problems such as requiring special hardware and taking a long processing time.

The present invention has been made in view of the above circumstances, and it is possible to record and store a large number of face images without increasing the memory capacity, and perform compressed data processing compatible with other systems. It is an object of the present invention to provide an ID photo system that can be used.

[0013]

According to the present invention, there is provided an image pickup means for picking up an ID face image, an area discriminating means for discriminating each area of a face image and a background image from an image picked up by this image pickup means,
A means for compressing partial data of the background image area discriminated by the area discriminating means to create representative compressed data, a data compression process is executed for the face image area discriminated by the area discriminating means, and the background image The area is composed of a means for replacing with the representative compressed data, and a storage means for generating and storing one screen of compressed image data from each data obtained by this means.

[0014]

As a result, according to the present invention, the data compression processing is executed only for the face image area discriminated from the image picked up by the image pickup means, and the background image area is preliminarily based on the partial data of the background image area. Since it can be replaced with the representative compressed data generated in step 1, the amount of data required for the background portion can be reduced and the compressed image data per screen can be minimized. Further, the compressed image data generated in this way can be applied to other systems, and the compressed image data of the face image can be used with compatibility.

[0015]

[Example] (First Example)

FIG. 1 shows a schematic configuration of the same embodiment.
In the figure, reference numeral 21 denotes a person who is a subject, and an image is taken in front of the background portion 22. The person 21 here is
An image of the upper half of the body is taken by the color TV camera 23.

In this case, the analog signal RGB which is the output signal from the color TV camera 23 is given to the AD converter 24. The AD converter 24 converts the analog signal RGB into a digital signal. Then, the digital data from the AD converter 24 is temporarily stored in the frame memory 25 as a frame image.

Then, the digital signal from the frame memory 25 is supplied to the DA converter 26 to be converted into an analog video signal, and this video signal is supplied to the output signal converter 27. The output signal converter 27 converts the video signal into, for example, an NTSC signal.

The output signal from the output signal converter 27 is transmitted to the TV.
A monitor 28 and a video printer 29 are provided. The TV monitor 28 is for visualizing a video signal as an image, and the video printer 29 is for printing out an image from the video signal.

On the other hand, 30 is a controller for controlling the entire system, and a keyboard 31 used by the user for inputting instructions is input to the controller 30.
Are connected. Further, a memory address control unit 32 is connected to the controller 30. The memory address control unit 32 performs address control when writing and reading the frame memory 25.

An image processing section 33 is connected to the frame memory 25. The image processing unit 33 cuts out a block image of 8 pixels × 8 pixels for image compression, a combination process, an RGB and YUV color conversion process, an image process for separating and extracting a background portion and a face image portion, and the like. , Various kinds of processing are performed.

A DCT processor 34, a quantizer 35, and an encoder 36 are connected to the image processor 33. Here, the DCT processing unit 34 performs processing such as discrete cosine transform and its inverse transform. The quantization unit 35 performs quantization or inverse quantization with a different step size for each DCT coefficient using a quantization table. The coding unit 36 performs variable length coding processing such as run length processing and Huffman coding performed by rearranging quantized data, or inverse coding processing thereof.

A face / background area extraction unit 37 is connected to the frame memory 25 and the image processing unit 33. The face / background portion area extraction unit 37 creates and records address map data regarding the background image area and the face image area of the upper half of the person from the image from the frame memory 25 based on the image processing in the image processing unit 33. I am trying.

An area discriminating section 38, a data memory 39, and an area-specific compressed data selecting section 40 are connected to the face / background area extracting section 37. Here, the area determination unit 38 compares the address map data of each area generated on the face / background area extraction unit 37 with the address of the compressed block image, and the area of this block image is the background. It is so arranged that it is determined whether it is a face part or a face part. The data memory 39 temporarily stores the quantized data of the DC-DCT coefficient in the background portion as representative compressed data. Whether or not the area-based compressed data selection unit 40 sets the processing data of the block image to be compressed according to the determination result of the background portion or the face portion by the area determination unit 38 as the background portion representative compressed data recorded in the data memory 39. , The DCT processing unit 34, the quantization unit 35, and the encoding unit 36 are used to selectively control the data. In this case, the switch unit 41 is controlled by the area-based compressed data selection unit 40 to select the side a. Alternatively, it can be switched to the b side.

A multiplexing unit 42 is connected to the switch unit 41. The multiplexing unit 42 multiplexes and decodes the compressed data of the face portion encoded through the encoding unit 36 or the representative compressed data of the background portion from the data memory 39.

Then, the external memory 4 is added to the multiplexer 42.
3 are connected. The external memory 43 stores and stores compressed data, CG image data for ID images, character data, etc. in a file, and is composed of, for example, a magneto-optical disk device or a hard disk device. Next, the operation of the embodiment will be described.

First, the color TV camera 23 captures an image of the upper half of the body including the face of the person 21 who is the subject. In this case, in order to make the person 21 easy to see, a monotone background portion 22 is provided behind the person 21 to capture an image. Further, it is assumed that the color TV camera 23 has previously been adjusted for white balance, lens aperture adjustment, and focus adjustment. Analog signal RG from color TV camera 23
B is sent to the AD converter 24, converted into a digital signal, and temporarily stored in the frame memory 25.

The digital signals stored in the frame memory 25 are sequentially read out in synchronism so that they can be output as video signals, sent to the DA converter 26, and converted into analog video signals. This video signal is sent to the TV monitor 28 and is displayed on the monitor as an upper half image of the person. Thereby, the lighting condition and the camera are adjusted so that the face image can be appropriately captured while looking at the monitor image.

From this state, in order to take a face image to be used as an ID image, the image data being taken is taken from the frame memory 2
5, a freeze control signal is given from the controller 30 to the frame memory 25 in order to freeze the face image on the frame memory 25. Such an operation is keyed in by the user from the keyboard 31 at the moment when the user wants to freeze. Then, the controller 30 stops the sequential write instruction of the frame memory 25 output from the memory address control unit 32, and this is realized. In this case, if the read data address control is continued, the image data in the frame memory 25 will be D regardless of the presence or absence of the freeze operation.
The image is output through the A converter 26 and displayed on the TV monitor 28 so that the camera image can be monitored. In addition, in order to print the ID card image, CG is previously stored in the external memory 43.
Data such as the logo image and name of the image are stored, and when printing, the data such as the logo image is transferred from the external memory 43 to the image processing unit 33, and the controller 30 controls the memory address control unit 32. Address control is performed so that a logo image or the like is written at a predetermined image position on the frame memory 25. Then, the face image of the compressed image is also read from the external memory 43, the image is expanded, and then written from the image processing unit 33 to a predetermined image position of the frame memory 25. And
When the ID card image is generated on the frame memory 25, it is sent to the video printer 29 by the controller 30 and printed out as an ID card image. Next, the main part of the embodiment will be described in more detail. First, after the image is set on the frame memory 25 as a still image, the image processing unit 3
In 3, the following operation is performed.

In this case, the image on the frame memory 25 is formed such that the background portion is A and the face portion of the person is B as shown in FIG. The absolute value of the difference in the pixel brightness (density) is calculated.

Now, assume that the data as shown in FIG. 3 is obtained on one line shown in FIG. As a result, in FIG. 3, the coordinate position where the difference value scanned from the left in the figure exceeds the threshold value t and the coordinate position where the difference value scanned from the right side in the figure greatly exceeds the threshold value t are the background portion A and the face, respectively. Boundary coordinate D of part B
Is obtained as. Then, by repeating such processing on each line from the top to the bottom in the vertical direction of the image, the background portion A
The face B is separated. In this way, in the image processing unit 33, both the background portion A and the face portion B are extracted by applying the predetermined threshold value t and performing the edge detection processing.

The result of the image processing unit 33 is given to the face / background area extraction unit 37. Then, the face / background portion area extraction unit 37 creates and records map data of each area address according to the areas of the background portion A and the face portion B extracted by the image processing unit 33.

Next, the block image of 8 pixels × 8 pixels in the upper left portion of the image shown in FIG. 2 is read out, and the area discrimination unit 38 is used to determine whether the address of the portion is the background portion A or the face portion B. The determination is made by referring to the map data of the part / background part area extraction part 37, and if it is the background part A, the data of that part is compressed and the representative compressed data is created.

In this case, the block data in the upper left portion of the image is read from the frame memory 25 and transferred to the image processing unit 33. Here, in order to perform data compression, first, the RGB data of the block image is converted into YUV data of luminance and other two color data. Then, the DCT processing unit 34
Each data is subjected to DCT (discrete cosine transform) processing. Then, the quantizer 3 is applied to the DC-DCT coefficient of each YUV.
Quantize to a different step size at 5. This quantized data is the representative compressed data here, and this representative compressed data is temporarily stored in the data memory 39. From this state, the compression of the image on the frame memory 25 starts.

In this case, the image shown in FIG. 2 is divided into blocks of 8 pixels × 8 pixels from the upper left end to the lower right end to perform compression processing. Here, the first block of 8 pixels × 8 pixels at the upper left end is the background portion A,
The controller 30 causes the face / background portion area extraction unit 37 and the area determination unit 38 to determine that the image block is the background portion A. Therefore, the switch unit 41 is connected to the side b by the region-specific compressed data selection unit 40. , The representative compressed data of the data memory 39 is transmitted to the multiplexing unit 42, and the external memory 4
3 will be recorded.

Next, regarding the processing of the block of 8 pixels × 8 pixels on the side of the block, since the background portion A is used here as well, the representative compressed data of the data memory 39 is multiplexed to that block as well, as described above. The external memory 43 is transmitted to the section 42.
Will be recorded in.

Thereafter, the processing is repeated for each block of 8 pixels × 8 pixels in the same manner. When the block to be processed is the face portion B, the face portion / background portion area extracting portion 37 and the area discriminating portion 38 cause the image to be processed. Since it is determined that the block is the face portion, the area-based compressed data selection unit 40 causes the switch unit 4
1 is connected to the a side.

In this case, when the 8 × 8 block image of the face B is transferred to the image processing unit 33, the RGB image is converted into YUV data of luminance and other two color data. Then, DCT processing is performed by the DCT processing unit 34 to obtain 8 × 8 DCT.
The quantization unit 35 quantizes the coefficient with a different step size for each coefficient, and the encoding unit 36 performs run-length processing and variable-length encoding processing, which is transmitted to the multiplexing unit 42, and is transmitted to the external memory. 43 will be recorded.

Next, if it is an 8 × 8 block image of the face portion B, the compression processing similar to the above is performed, and thereafter, these processings are repeatedly executed, and if the background portion A is judged to be the background portion, the representative compression is performed. The data is recorded in the external memory 43 via the multiplexing unit 42, and when the face portion B is determined to be the face portion, compression processing is performed, and this is recorded in the external memory 43 via the multiplexing unit 42. Then, the compressed data for one screen including such a face image is recorded together with the file name for identifying the ID number of the person involved in photographing. Next, the restoration by decompressing the compressed image will be described in detail.

Decompression of this compressed image is the reverse process of compression. In this case, the file name data of the image to be expanded is input from the keyboard 31 and transmitted to the controller 30. Then, the corresponding file image is read from the external memory 43.

When the compressed data of the file is input to the multiplexing unit 42, it is demultiplexed. in this case,
The area-based compressed data selection unit 40 causes the switch unit 41 to
If it is connected to the side, the compressed data is transferred to the encoding unit 36 and inversely encoded, is returned to a block image of 8 pixels × 8 pixels, and is inversely quantized by the quantization unit 35 to be close to the original value. DC
It is returned to the T coefficient. Further, the DCT processing unit 34 performs an inverse DCT operation to generate YUV image data of 8 pixels × 8 pixels, transfers the YUV image data to the image processing unit 33, converts the YUV image data into RGB image data, and restores one block. Then, the image data of this one block is transferred to the memory address control unit 3
Frame memory 25 by address control of 2
Write in. By performing such decompression processing for all blocks, a face image on one screen can be displayed in the frame memory 25.
Will be restored on top.

Here, in order to make the connection between the blocks smooth, the image processing unit 33 to the frame memory 2 is used.
If the smoothing process is performed by the image processing unit 33 before transferring to 5, an image that is easy to see can be obtained.

Therefore, in this way, the area of the background image and the face image area of the frame image of the frame memory are separated, and the compressed data is changed in the background image and the face image area and recorded and saved. That is, general compression processing is performed in the face image portion, and certain color data in the background portion is compressed in advance to generate representative compressed data for the background portion, and this representative compressed data is compressed in the background portion. Since it is used as a result, the amount of data required for the background portion can be significantly reduced, and a large number of ID face images can be recorded and saved without increasing the memory capacity. Also, the compressed image data created by this system can be applied to other systems, and the compressed image data created by other systems can also be applied to this system. Image data can be used. Further, since the background portion can be image data of one type of color, the contrast of the background image can be eliminated, and an ID image with a neat appearance can be obtained. (Second Embodiment) FIG. 4 shows a schematic configuration of the second embodiment, and shows only a portion different from FIG. 1 described above.

In this case, DC-DC in the background portion described above
A DC of a predetermined color C is stored in the data memory unit 39 that temporarily stores the data after the quantization of the T coefficient as the representative compressed data.
A second data memory 4 for obtaining compressed data after quantization of DCT coefficients and storing it as second representative compressed data.
5 and a third data memory 46 in which compressed data after quantization of the DC-DCT coefficient of the color D different from the destination used for image expansion is obtained in advance and stored as third representative compressed data, Further, the representative data discriminating unit 47 for discriminating whether or not the data processed after the inverse encoding at the time of image expansion is the background portion and the second representative compressed data of the color C, the representative data discriminating unit 47. Compressed data selection for converting the second representative compressed data of the color C into the third representative compressed data of the color D in order to change the discrimination signal from the third representative compressed data of the color D It has a section 48. Others are the same as those in FIG. 1, and the same portions are denoted by the same reference numerals. Next, the operation of the embodiment will be described.

The basic operation is in accordance with the first embodiment, but here representative compressed data is created in advance,
It is stored in the second data memory 45. Therefore, the second
The data memory 45 may be implemented by a ROM.

If the image block to be processed at the time of compression is the background part, the area-specific compressed data selection section 40 connects the switch section 41 to the b side to combine the representative compressed data of the second data memory 45 with the multiplexing section. Tell 42. If the image block is a face portion, the same process as above is performed. In this case, the switch unit 49 by the compressed data selection unit 48 is c
It is connected to the side so that data can be transferred from the quantizer 35 to the encoder 36. As a result, the face image having the background color changed to C is recorded in the external memory 43 via the multiplexing unit 42 together with the file name for identifying the ID number of the person who has taken the image. On the other hand, in the case of decompression, the same as in the first embodiment, but the data of the background part is the color C
In that case, the difference is that the face image is created with color D.

In this case, when decompressing file data is input from the keyboard 31 and transmitted to the controller 30, the corresponding file image is read from the external memory 43.

When the compressed data of the file is input to the multiplexing unit 42, it is demultiplexed. in this case,
The area-based compressed data selection unit 40 causes the switch unit 41 to
If it is connected to the side, the compressed data is transferred to the encoding unit 36, is inversely encoded, and is returned to a block image of 8 pixels × 8 pixels. At this time, the AC-DCT coefficient is zero and DC-DC
The representative data discriminating unit 47 determines whether the data value corresponding to the T coefficient is the second representative compressed data at the time of compression,
If they match, the compressed data selection unit 48 receives the determination signal and connects the switch unit 49 to the d side.
Then, the value of the third representative compressed data from the third data memory 46 is transmitted to the quantizer 35, and the inverse quantization process is performed. If they do not match, the compression data selection unit 48 keeps the switch unit 49 connected to the c side and the quantization unit 3
Tell 5.

In the quantizer 35, inverse quantization is performed and the DCT coefficient is returned to the original value of the block of 8 pixels × 8 pixels. Further, this is subjected to inverse DCT calculation in the DCT processing unit 34 to generate YUV image data of 8 pixels × 8 pixels, transferred to the image processing unit 33, and converted into RGB image data to restore one block. Then, the image data of one block is sequentially written in the frame memory 25 by the address control of the memory address control unit 32.

A face image of one screen is restored in the frame memory 25 by performing such expansion processing for all blocks. The color of the background portion in this case is D. Also in this case, an image that is easy to see can be obtained by performing smoothing processing in the image processing unit 33 before transferring from the image processing unit 33 to the frame memory 25 in order to smooth the connection between blocks.

Therefore, in this way, the color of the background portion is set to the predetermined color C at the time of compression processing, and it is judged at the time of expansion whether the compressed data is the predetermined color C. Can be replaced with the color D to be expanded, and the color of the background portion of the face image once photographed and created can be changed. This means that if an image compressed with this system is used in another system, it will be restored with the background color when the image was compressed, but if the image was expanded with the original system, a card image will be printed. In this case, the background color can be changed and restored. For example, in the application of issuing an ID card employee ID card, the background color can be changed for each affiliation, business office, or the like. (Third embodiment)

By the way, when a face image that has been once photographed and created and saved is reproduced and finished as an ID card, the printed image is often a person's position deviated to the left or right, which is very attractive. It was often a bad ID card.

This is because the center of the monitor TV to be displayed and the center of the captured image may not necessarily coincide with each other and may be photographed with a bias. Also, at the time of photographing, the person's position may become the center of the image. It was also troublesome to operate the camera.

Therefore, in the present embodiment, if a person is made to come to the center of the monitor TV to some extent when the face image is taken, then the face image is automatically corrected so that the position of the person is at the center of the screen. A face image for an ID card that does not fail can be recorded and saved.

FIG. 5 shows a schematic configuration for realizing the same embodiment. In the figure, 51 is a person who is a subject, and an image is taken in front of the monotone background portion 52. The person 51 has a color TV camera 53 on the upper body.
I am trying to take an image.

In this case, the analog signal RGB which is the output signal from the color TV camera 53 is given to the AD converter 54. The AD converter 54 converts the analog signal RGB into a digital signal. Then, the digital data from the AD converter 54 is temporarily stored in the frame memory 55 as a frame image.

Then, the digital signal from the frame memory 55 is given to the DA converter 56 to be converted into an analog video signal, and this video signal is given to the output signal converting section 57. The output signal converter 57 converts the video signal into, for example, an NTSC signal.

The output signal from the output signal converter 57 is sent to the TV.
A monitor 58 and a video printer 59 are provided. The TV monitor 58 is for visualizing a video signal as an image, and the video printer 59 is for printing out an image from the video signal.

On the other hand, 60 is a controller for controlling the entire system, and a keyboard 61 used by the user for inputting instructions is input to the controller 60.
Are connected. The keyboard 61 here is also used for inputting weight values and various instructions to be described later.

In the frame memory 55, the skin color detecting section 6
2. Skin color area centroid position coordinate calculation unit 63, background color detection unit 6
4. The background color area centroid position coordinate calculation unit 65 is connected. Here, the flesh color detecting unit 62 inputs, for example, a color data value, which is a standard Japanese flesh color, from the keyboard 61 in advance, and detects a color considered to be a flesh color from the image based on the color data value. I am trying. The skin color area barycentric position coordinate calculation unit 63 obtains an image area of a color considered to be a skin color detected by the skin color detection unit 62 and calculates barycentric position coordinates of the area. The background color detection unit 64 detects, for example, a color of a portion (upper right end or upper left end of the image) which is apparently a background color, obtains a representative background color from the color data by the controller 60, and uses that color as a reference. The image is detected whether it is within the range of the color. The background color area barycentric position coordinate calculation unit 65 obtains a color image area of a color considered to be the background color detected by the background color detection unit 64 and calculates the barycentric position coordinates of the area.

Then, the skin color area barycenter position coordinate calculation unit 63
A person image center position coordinate calculation unit 66 is connected to the background color area barycenter position coordinate calculation unit 65. The person image center position coordinate calculation unit 66 calculates the center position coordinates of the person image from the center of gravity position coordinates of the respective regions calculated by the skin color region center of gravity position coordinate calculation unit 63 and the background color region center of gravity position coordinate calculation unit 65. Is. In this case, in the calculation in the person image center position coordinate calculation unit 66, the weighting coefficient Wa 67 calculated from the face portion (or input from the keyboard 61) and the weighting coefficient Wb 68 calculated from the background portion (or input from the keyboard 61). Is being used.

A person image cutting-out section 69 is connected to the person image center position coordinate calculating section 66. The person image cutout unit 69 cuts out the person image from the frame memory 55 based on the center position coordinates of the person image calculated by the person image center position coordinate calculation unit 66, and records and stores the image data in the external memory 70. ing. The external memory 70 here comprises a magneto-optical disk, a hard disk, or the like. Next, the operation of the embodiment will be described.
In this case, the color TV camera 53 is used as a still video camera to photograph the upper half of the person 51. The analog signal RGB from the color TV camera 53 is converted into an AD converter 54.
To the frame memory 5
5 is temporarily stored.

Then, the digital signals stored in the frame memory 55 are sequentially read out in synchronization with each other so that they can be output as a video signal, supplied to the DA converter 56, and further supplied to the output signal converter 57. NTSC
It is output as a signal. This signal is sent to the video printer 59.
If you input to, you can output the print as an image,
When input to the TV monitor 58, the image on the frame memory 55 can be monitored.

In this case, in order to record an image of one screen taken by the color TV camera 53 as a still video camera in the frame memory 55, the video signals repeatedly sent from the color TV camera 53 are frame-by-frame in the frame memory 55. , The controller 60 controls the write address of the frame memory 55 in synchronization with the video signal.

Then, in this way, the frame memory 5
For the image recorded on the image No. 5, the flesh color detecting section 62 detects the flesh color and the background color detecting section 64 detects the color of the background section according to the address control of the controller 60.

In this case, the flesh color detecting unit 62 determines, for example, a color considered to be a standard Japanese flesh color (R0, G0, B0).
As skin color (R0 ± Δr, G0 ± Δg, B0 ± Δb)
Is set (these color data values are input from the keyboard 61 in advance), and all the color points in this color range are detected.

Further, the background color detecting section 64 samples several colors of a portion (for example, the upper right end or the upper left end) which is apparently the background, and the controller 60 calculates the average value of these color data values. The color (R0 ', G0
′, B0 ′), and based on this color, the background color is (R0 ′ ± Δr ′, G0 ′ ± Δg ′, B0 ′ ± Δb ′)
Is set, and all color points within this color range are detected.

When the histogram of each color of the face image is taken, it seems that the most color points in the background portion are extracted. Therefore, the histogram is calculated by the controller 60.
And the color is (R0 ', G0', B0 '), and the range is (R0' ± Δr ', G0' ± Δg ', B0' ± Δb).
You may make it detect all the color points in ′). Further, in the above, the color detection is performed in the RGB color space,
For example, it is possible to detect a color using a hue in the LUV color space or the like.

The detection result of the skin color detecting section 62 is sent to the skin color area barycentric position coordinate calculating section 63. In this case, the skin color center of gravity position coordinates calculation unit 63 calculates the skin color center of gravity position coordinates (x
The calculation of a, ya) is performed as follows.

Here, if the total number of dots of the color points in the image area of the skin color detected by the skin color detection unit 62 is P, and the number of the skin color points on one line in the vertical direction is px,
With the x coordinate on that line as xa, xa can be calculated by xa = Σ (px.x) / P. On the other hand, the same is done for ya on the y coordinate.

This is the calculation of the barycentric position coordinates (xa, ya) of the skin color image area. The barycentric position coordinates (xb, yb) of the background color image area of the background portion in the background color area barycentric position coordinate calculating unit 65 are calculated. ) Is also calculated in the same manner as described above.

Next, the weight coefficient Wa 67 and the weight coefficient Wb 6
8 is obtained as follows. In this case, the skin color image area 7 in the face image read by the skin color detecting unit 62 while the image data 70 in the frame memory 55 as shown in FIG.
Let Pa be the total number of dots of color points within 1 and Pb be the total number of dots of color points within the background color image area 72 of the background portion read by the background color detection unit 64. Send Pa and Pb to controller 60.

In the controller 60, the quotient obtained by dividing Pb by Pa is calculated as C, and the value is temporarily stored in the controller 60. The controller 6 includes a keyboard 6 in advance.
There is the threshold value D input in 1, and the controller 60 compares C and D. At this time, if 0 <C ≦ D, Wa>
Wb, and if C> D, then Wa <Wb, and the weighting factor W
The controller 60 calculates a and Wb. Then, the calculated Wa and Wb are transferred to the weight coefficient Wa 67 and the weight coefficient Wb 68, respectively. The formula used here is as follows. Wa = D * Pa / (D * Pa + Pb) Wb = Pb / (D * Pa + Pb)

The controller 60 address-controls the image data in the frame memory 55 to count the total number of pixels, and the controller 60 calculates Pc by dividing Pb by Pc and E as the quotient. Alternatively, Wa and Wb may be calculated by the controller 60 in the same manner as described above. Here, the former method is used. Next, the position coordinate (xc, yc) of the center of the person image is calculated by the person image center position coordinate calculation unit 66.

In this case, the horizontal width L of the image captured from the TV camera 51 is calculated. As shown in FIG. 7, the coordinates of the upper left corner and the lower right corner of the image photographed for convenience are set to (x0, y0) and (x1, y1). Further, the captured image is made to correspond to an address on the frame memory 55, and the first horizontal line of the image data is scanned while the address is controlled by the controller 60 to detect the presence or absence of the image data. At that time, the coordinate value of the image data detected first is set to (x0, y0), and the value is input to the person image center position coordinate calculating section 66. Also,
With the first coordinate value at which one line is scanned and image data cannot be detected as x1 ', the controller 60 calculates the value x1 of the x coordinate of the lower right corner by the following formula. Then, the coordinate values are input to the person image center position coordinate calculation unit 66. x1 = x1'-1

On the other hand, the value y1 of the y coordinate of the lower right corner is calculated by
After the x-coordinate value x1 of the lower right corner is calculated, it is performed as follows. The address of the captured image data is controlled by the controller 60 in correspondence with the address of the frame memory 55, and one line is scanned in the vertical direction from the coordinates of the image data detected first, and the first image data cannot be detected. With the coordinate value as y1 ′, the controller 60 calculates the y coordinate value y1 at the lower right corner by the following formula. Then, the coordinate values are input to the person image center position coordinate calculation unit 66, y1 = y1'-1

These two input coordinate values (x0, y0
) (X1, y1) to the human image center position coordinate calculation unit 6
In step 6, the horizontal width L is calculated. The formula is as follows: L = x1−x0 + 1 The width of the image in the vertical direction is calculated in the same manner.

Next, the position of the center of the portrait (xc, yc)
To calculate. The center of gravity (xa, xa,
As shown in FIG. 8, ya) (xb, yb) generally has a positional relationship of coexisting about kk 'when the center line in the vertical direction of the captured image is kk'. Therefore, the controller 60 is first sent to the coordinate data value (xb, yb
) Is transferred, and calculation is performed to move symmetrically around kk 'to the side having (xa, ya). The coordinate value is (x
b ′, yb ′), and the coordinate data values are input again to the person image center position coordinate calculation unit 66. Here, xb 'is calculated by the following formula. xb '= L-xb

Then, by substituting the coordinate values (xb ', yb') and (xa, ya) into the following equation,
The position coordinate (xc, yc) of the center of the person image is calculated by the person image center position coordinate calculation unit 66, and the coordinate value is transferred to the person image cutout unit 69. xc = (Wa * xa + Wb * (L-xb ')) / (Wa + Wb) Also, yc is calculated in the same manner as described above. However, yc
In the calculation of w, it is necessary to make wb negative so as not to cut the hair part of the face screen. Next, the clipping of the image will be described. Here, the case where the face image corresponding to the skin color area 71 in the image data 70 shown in FIG. 9 in the frame memory 55 is cut out will be taken up.

The position coordinates (xc, yc) of the center of the person image calculated by the person image center position coordinate calculating section 66 are sent to the person image cutting section 69. Portrait image cropping unit 69
Then, a command for cutting out the image area by the same L '/ 2 width in the x direction and the y direction centering on these coordinate values is sent to the frame memory 55. In this case, the vertical width is L '= y
1-y0 + 1. Then, the address of the frame memory 55 is made to correspond to the image area, and the data is taken out from the frame memory 55 while controlling the addresses of y0 from xc- (L '/ 2) to xc + (L' / 2).

Next, from xc- (L '/ 2) to xc + (L
Up to '/ 2), data is taken out from the frame memory 55 while controlling the address for y1 from the line next to y0. The person image data extracted in this way is sent to the external memory 70 and recorded and stored.

Therefore, in this way, the position coordinates of the center of the person image are calculated by multiplying the respective gravity centers by the weighting factors from the calculated positional relationship of the gravity centers of the skin color area of the face image and the background color area of the background portion. Is calculated and the person image is cut out. Thereby, for example, the weighting factor is Wa =
When 0.6 and WB = 0.4, the position coordinates of the center of the human image can be calculated in consideration of the displacement of the face position that occurs when the hair portion of the face image is offset to the left or right.

Since the center coordinates of the human image are calculated by multiplying the center of gravity of the two color areas by the weighting factor as described above, for example, if the weighting factor is Wa = 1 and WB = 0. The center of gravity of the skin color area of the face image may be processed as the center coordinates of the person image.

The horizontal width and the vertical width of the image to be cut may be set to different widths depending on the case, and the cut-out may be performed. That is, it is possible to cut out not only a square shape but also a vertically or horizontally long rectangular area. Further, as described above, the weighting factor is automatically calculated from the captured image and the center position coordinates of the human image are calculated, but the weighting factor may be set to an arbitrary value for calculation processing. In that case, the weighting factor W is set by the keyboard 61.
By inputting a and Wb, the center position coordinates of the person image can be calculated.

As a result, when the face image is photographed so that the person to be photographed is located in the middle of the monitor TV to some extent, after that, the face screen position is automatically corrected so as to come to the center of the screen. A face image for an ID card that does not fail can be generated and recorded and saved. Further, it is possible to eliminate the displacement of the face position that occurs when the hair portion of the face screen is biased to the left or right.

[0086]

According to the present invention, the data compression processing is executed only for the face image area discriminated from the image picked up by the image pickup means, and the background image area is preliminarily based on the partial data of the background image area. Since it can be replaced with the generated representative compressed data, the amount of data required for the background portion can be reduced and the compressed image data per screen can be minimized. As a result, a large number of ID face images can be stored in the memory. Records can be saved without increasing. Further, since the compressed image data thus generated can be applied to other systems, it becomes possible to use the compressed image data of the face image with compatibility.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of a first embodiment of the present invention.

FIG. 2 is a diagram for explaining the first embodiment.

FIG. 3 is a diagram for explaining the first embodiment.

FIG. 4 is a diagram showing a schematic configuration of a second embodiment of the present invention.

FIG. 5 is a diagram showing a schematic configuration of a third embodiment of the present invention.

FIG. 6 is a diagram for explaining a third embodiment.

FIG. 7 is a diagram for explaining the third embodiment.

FIG. 8 is a diagram for explaining the third embodiment.

FIG. 9 is a diagram for explaining the third embodiment.

FIG. 10 is a diagram for explaining a conventional example.

FIG. 11 is a diagram for explaining another conventional example.

[Explanation of symbols]

21, 51 ... Person, 22, 52 ... Background part, 23, 53 ...
Color TV camera, 24, 54 ... AD converter, 25, 5
5 ... Frame memory, 26, 56 ... DA converter, 27,
57 ... Output signal converter, 28, 58 ... TV monitor, 2
9, 59 ... Video printer, 30, 60 ... Controller, 31, 61 ... Keyboard, 32 ... Memory address control section, 33 ... Image processing section, 34 ... DCT processing section, 35 ... Quantization section, 36 ... Encoding section, 37 ... Face / background area extraction section, 38 ... Area discrimination section, 39 ... Data memory, 40 ... Area-specific compressed data selection section, 41 ... Switch section, 42 ... Multiplexing section, 43 ... External memory, 45 ... 2 data memory, 46 ... Third data memory, 47 ... Representative data discriminating section, 48 ... Compressed data selecting section, 49 ... Switch section, 62 ... Skin color detecting section, 63 ... Skin color area barycentric position coordinate calculating section, 64 ... Background color detection unit, 65 ... Background color area barycentric position coordinate calculation unit, 66 ... Portrait image center position coordinate calculation unit, 6
7 ... Weighting factor Wa, 68 ... Weighting factor Wb, 69 ... Portrait image cutout unit, 70 ... External memory.

Claims (1)

[Claims] 1. An image pickup means for picking up an ID face image, an area discriminating means for discriminating each area of the face image and the background image from the image picked up by the image pickup means, and a background discriminated by the area discriminating means. A means for compressing partial data of the image area to create representative compressed data; and a data compression process for the face image area discriminated by the area discriminating means and replacing the background image area with the representative compressed data. An ID photo system comprising: a means and a storage means for generating and storing one screen of compressed image data from each data obtained by the means.