JPH06350914A - Pink-eye effect correction system - Google Patents

Abstract

PURPOSE:To correct an image including a pink-eye effect to a normal eye by converting the image including the pink-eye effect into digital picture data thereby detecting automatically the pink-eye effect. CONSTITUTION:A scanner 19 is used to read an image of a film 4 and the image is displayed on a monitor 14 so as to allow the user to visually check the image of the film 4 before its print. A position in the vicinity of a pink-eye effect is pointed out in a frame discriminated by the user that the frame includes the pink-eye effect as the result of visual inspection on the monitor 14 and a work station 10 discriminates automatically the pink-eye effect from the image of the in the vicinity to be pointed out and corrects the pink-eye effect automatically.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a red-eye correction system for detecting a red-eye generated when a subject such as a person is stroboscopically photographed and automatically restoring the red-eye.

[0002]

2. Description of the Related Art Conventionally, the so-called red-eye phenomenon, in which the human eye glows red or gold, has been a problem in stroboscopic photography. The red-eye phenomenon occurs when strobe light that has passed through the pupil of the eye is reflected by the retina and the reflected light is reflected on the film.

And, in order to solve this problem conventionally,
For example, in Japanese Unexamined Patent Publication No. 2-64532, it is determined that the subject has a low luminance and the shooting condition is such that a red eye occurs when shooting with an electronic flash device, and red eye information indicating that is recorded in a margin of the film. On the printer side, however, a technique relating to "a camera capable of recording shooting information and an information output device for visualizing subject information" is disclosed, in which a frame for which red-eye information is detected is subjected to correction processing after printing.

Further, Japanese Patent Application Laid-Open No. 2-114253 discloses a technique relating to a "printer device" which adjusts the amount of printing light applied to the red-eye portion of a photographic original during printing exposure.

[0005]

However, in the technique disclosed in the above-mentioned Japanese Patent Laid-Open No. 2-64532,
Although the correction process is described in detail, since the method of adding and correcting after printing is adopted, the correction takes time and requires a specialized technique.

Further, according to the technique disclosed in Japanese Patent Laid-Open No. 2-114253, it is generally difficult to correct the color of only the red-eye portion because the red-eye portion is a small area with respect to the entire area of the photograph. Is.

The present invention has been made in view of the above problems, and an object thereof is to convert an image containing red eyes into digital image data, automatically detect the red eyes, and restore the normal eyes. It is in.

[0008]

In order to achieve the above object, a red eye correcting system according to the first aspect of the present invention comprises a scanner means for converting an image of a negative color film into digital image data, and a scanner means for converting the digital image data. It is characterized by comprising red eye extracting means for extracting red eyes from the inside, and red eye correcting means for correcting the red eyes extracted by the red eye extracting means.

The red-eye correcting system according to the second aspect is a monitor means for visually inspecting an image of a negative color film before printing, and a red-eye vicinity in a frame which is judged to be red-eye as a result of the visual inspection by the monitor means. A designating means for designating the position of, and a red eye discriminating means for automatically discriminating a red eye from the image in the vicinity designated by the designating means,
And a red eye correcting means for correcting the red eye discriminated by the red eye discriminating means.

Further, the red-eye correction system according to the third aspect is a recording means for recording information on the possibility of occurrence of red-eye on a negative color film, and a reading means for reading the information recorded on the negative color film by the recording means. A monitor means for displaying the image of the negative color film before printing, an instruction means for manually instructing the position near the red eye with respect to the frame in which the red eye may occur, and an instruction means for instructing the position. The present invention is characterized by comprising red-eye determining means for automatically determining red-eye from a nearby image and red-eye correcting means for correcting the red-eye determined by the red-eye determining means.

[0011]

That is, in the red eye correcting system according to the first aspect of the present invention, the scanner means converts the image of the negative color film into digital image data, and the red eye extracting means extracts the red eye from the digital image data, The correcting means corrects the red eye extracted by the red eye extracting means.

In the red-eye correction system according to the second aspect, the monitor means visually inspects the image of the negative color film before printing, and the designating means is the red eye as a result of the visual inspection by the monitor means. To indicate the position in the vicinity of the red eye, the red eye discriminating means automatically discriminates the red eye from the image in the vicinity indicated by the designating means, and the red eye correcting means corrects the red eye discriminated by the red eye discriminating means. .

Further, in the red-eye correction system according to the third aspect, the recording means records the information on the possibility of the occurrence of red eye on the negative color film, and the reading means records the information recorded on the negative color film by the recording means. The reading / monitoring means displays the image of the negative color film before printing, the instructing means manually instructs the position near the red eye to the frame in which the red eye may occur, and the red eye discriminating means uses the instructing means. The red eye is automatically discriminated from the image in the vicinity indicated by the red eye discriminating means and the red eye correcting means corrects the red eye discriminated by the red eye discriminating means.

[0014]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing the configuration of a red-eye correction system according to the first embodiment of the present invention. In FIG. 1, white light from the light source 1 is guided to the color line sensor 6 via the diffuser 2, the film 4 in the film carrier 3, and the lens 5. Therefore, the image of the film 4 is formed on the color line sensor 6 by the lens 5. Further, the output signal of the color line sensor 6 is converted into a digital amount by the A / D converter 8 via the amplifier circuit 7. Then, the signal converted into the digital amount is sent to the workstation 10 via the I / O boat 9 and stored in the memory 11 as image data. Then, the sensor scanning device 12 scans the color line sensor 6 in a direction orthogonal to the sensor 6 to obtain image data for one frame. And video RAM
Data for monitor display, which is obtained by converting the image data into a predetermined density and color, is stored in 13, and the display data is displayed by the monitor 14.

Further, the magnetic head 15 reads the data recorded in the magnetic recording portion of the film 4, as will be described later. The magnetic data read by the magnetic head 15 is amplified and waveform-shaped by the signal processing circuit 16, sent to the workstation 10 via the I / O port 9, and stored in the memory 11. Reference numeral 17 is a film feed roller for feeding the film 4. Also,
Reference numeral 18 is a notch detection sensor. In this embodiment, the scanner 19 is configured by the above-described units.

In addition, correction data, the number of notches, interval data within each notch, and the like are recorded in the LSI card 20.
Then, the workstation 10 performs various calculations and controls such as print color balance control and density control described later based on various information obtained from the scanner 19. Further, the printer 21 prints an optimum print based on the information calculated by the workstation 10.

Next, FIG. 2 is an overall control block diagram of the camera. In FIG. 2, when the exposure is started, the photometric sensor 23
The output analog signal is amplified by the amplifier circuit 24 at the subsequent stage and further converted into a digital signal by the A / D converter 25.

The photometric data converted into the digital value is further stored in a predetermined storage area of the memory in the CPU 22. Further, in the CPU 22, an apex operation is executed between this photometric data, the film sensitivity SU and the photographing lens aperture value AU and the exposure correction value CU input by the photographing information input circuit 26, and the shutter speed TV is calculated in this way. It

Further, the distance measuring principle of the distance measuring optical system 27 is obtained from the distance between two images obtained by dividing the pupil of a known photographing lens, and the distance measuring optical system 27 is a field lens separator. It is composed of a lens and the like. The output signal of the CCD 28, which is sequentially sent according to the timing signal of the drive circuit 29, is amplified by the amplifier circuit 30 and then converted into a digital value by the A / D converter 31. Further, the CCD 2 converted into this digital value
The output signal of 8 is stored in the memory in the CPU 22.
Then, the CPU 22 calculates the defocus amount of the photographing lens based on the output signal of the CCD 28. Further, the distance 1 to the subject is also calculated.

Then, the CPU 22 drives the lens drive circuit 32 based on the defocus amount to set the photographing lens to the straight point position. In addition, the above-mentioned direct point detection operation is performed by the AF.
It is performed after the CPU 22 detects that the start switch 43 is closed. Furthermore, from the CPU 22 to the strobe device 3
The signal S1 is output for the signal No.3. This S1 is a signal for causing the strobe device 33 to start strobe emission, and whether or not to emit strobe light is controlled according to the result of the apex operation. Further, the liquid crystal display unit 34 displays the photographing information such as the shutter speed and the mode of the camera, and the liquid crystal drive circuit 35 controls the driving thereof.

Further, when the strobe 33 emits light and the distance to the subject is shorter than a predetermined value, it is considered that a red eye is likely to occur, and the magnetic recording section 37 of the predetermined section of the film 36 is passed through the magnetic head 39. Record that information.
At this time, the magnetic head 39 is driven and controlled by the signal processing circuit 38. Incidentally, FIG. 3 is a plan view of the film 36 having the magnetic recording portion 37. Further, when the release switch 40 is closed, the power supply to the shutter front curtain locking magnet 41 is cut off, and the shutter front curtain starts traveling. When the film exposure is performed for a predetermined time, the power supply to the shutter rear curtain locking magnet 42 is cut off, the shutter rear curtain travels, and the exposure is completed.

The operation of the first embodiment will be described below with reference to the flow chart of FIG. In the first embodiment, when the operation is started, shading correction is first performed. In this correction, the distortion due to the sensitivity variation of the color line sensor 6 and the uneven light amount of the light source 1 is corrected (step S).
1). Next, based on the image data detected by the color line sensor 6, the average transmission density (LA
TD) is measured (step S2). And LAT
Monitor 1 from D and correction data from LSI card 20
4 to display the image of film 4 on video data R,
G and B densities and color corrections are performed (steps S3 to S).
5). The corrected video data is then stored in the video RAM1.
3 is stored (step S6). Next, the magnetic data of the magnetic recording portion 37 of the film 4 is read by the magnetic head 15 (step S7). As a result, when there is a possibility that redeye occurs in the frame currently being processed, redeye correction is executed (step S8, 9). This red-eye correction program will be described in detail later.

If there is no possibility of red-eye generation in step S8, then the print density and R, G, B color corrections are performed (step S10). Then, after the reflectance-voltage linear signal (BGR) is converted into a density-voltage linear signal (YMC) by a γ correction unit (not shown), the gray components in the three color signals are separated (undercolor removal; UCR). Is called)
Then, a black signal is generated (steps S10 to S13).
Then, color correction masking is performed using a predetermined masking equation (step S14).

Next, after gradation correction is performed (step S
15), edge enhancement (step S16), and sharpness correction such as smoothing is performed (step S17).
Next, printing is executed based on the image data subjected to the above processing (step S17). And film 4
Is sent by one frame and the same processing is executed again (step S18).

Next, the sequence of the subroutine "red eye correction" will be described with reference to the flowchart of FIG.
First, noise removal by a filter is performed (step S20). Next, the red-eye candidate pixels are labeled while sequentially processing the image data from the upper left to the lower right of the image (step S21). Here, only pixels of a color that may cause red eye are extracted. That is, with C _{1 to} C ₄ as constants, pixels having colors of C ₁ <G / R <C ₂ and C ₃ <B / R <C ₄ are selected as red eye candidates.

Subsequently, the red-eye contour line is extracted from the red-eye candidate pixel again while sequentially processing from the upper left to the lower right of the image (step S22). Here, a known contour line extraction method will be described with reference to FIG. A 2 × 2 mask is used for extracting the contour line in this embodiment, as shown in FIG. Put your left hand on the wall and proceed,
In addition, when you return to the original position again, the path that you followed becomes a contour line. That is, in this method, the point on the contour line which is currently obtained is continuously placed at the lower point on the left side in the traveling direction in the 2 × 2 mask shown in FIG. 4A, that is, the position of x _{k in} the figure. To do. Then, depending on whether or not the points a and b in the traveling direction in the figure are red-eye colors, the next x _{k + 1} >
Determine y _{k + 1} .

Then, as shown in FIG. 4 (b), when a is not red-eye color and b is not red-eye color, it is 1 in the traveling direction as it is.
Going forward, let a be the next x _{k + 1} and b be the y _{k + 1} . Also,
a, b are both when the red-eye color, new x _{k + 1} by rotating the traveling direction 90 degrees right, as shown in FIG. 4 (c) position of b, y k _{+ 1} is the raw y _k positions To do. At this time, the new a and b are the positions of a'and b'in the figure. Also, a
Is not red-eye color, the direction of travel is rotated 90 degrees to the left regardless of whether or not b is red-eye color, and x _{k + 1} remains x _k
Then, y _{k + 1} is set to the position of a. This operation starts from a certain starting point, and stops after one lap. In the above contour line extraction, the coordinates are labeled for the red-eye pixels (see FIG. 5A). That is, with the starting point as (0, 0), the coordinates are changed each time the process advances or moves backward in the x direction and the y direction.

When the contour line extraction is completed by such a method, a subroutine "red eye recognition" is executed in step S23. This detailed operation will be described later. Then, in step S24, when FLG = 1 is not satisfied, the red-eye candidate pixel labeling is corrected (step S28). This is a work for avoiding the same work of extracting the contour line again because the contour line is extracted for all the pixels in the lower right pixel while scanning from the upper left of the image. For example, if the red-eye candidate pixel is set to "1", the above "1" may be changed to "0" unless FLG = 1 in step S24.

When FLG = 1 in step S24, the iris edge is detected next. This is shown in FIG.
6B, the density distribution of pixels on a horizontal line passing through the center of gravity (x ₀ , y ₀ ) of the red-eye contour line is detected, and iris edges x ₁ and x ₄ as shown in FIG. 6B are detected. Find out if In FIG. 6B, P ₁ is a cornea reflection image (Purkinje image) of stroboscopic light emission. If no iris edge is detected, the red-eye candidate pixel labeling is corrected in the same manner as above.

Subsequently, if an iris edge is detected in step S26, the color inside the detected red-eye contour line is corrected (step S27). Specifically, the color is changed to black.
The above operation is sequentially repeated from the upper left pixel to the lower right pixel of the image (steps S29 and S30).

Next, the operation of the subroutine "red eye recognition" will be described with reference to the flowchart of FIG. When this subroutine is executed, first, FIG.
The center of gravity O (x ₀ , y ₀ ) of the red-eye contour line as shown in FIG. 5B is obtained based on the coordinates of the contour line shown in FIG.
31). Next, the distances r _{0 to} r ₇ to the contour lines of the 8-direction vector starting from the center of gravity are obtained (step S3
2). Then, the average value r of r _{0 to} r ₇ is obtained (step S33), and this r _{0 to} r ₇ is normalized by r and this is r
₀ 'to r _7' and (step S34), r ₀ 'to r
Obtaining the minimum value rmin 'and' maximum rmax of _'7 (step S35).

Next, when the average value r is not C ₁ <r <C _{2 with} respect to the predetermined constants C ₁ and C ₂ , the red eye is very small and the red eye is not noticeable, or the red eye is abnormally large and the red eye is possible. The red-eye recognition flag FLG is also set to "0" to indicate that there is no nature, and the process returns from the subroutine (steps S36, 39,
40).

Then, when C ₁ <r <C ₂ , then C
_{When 3} <rmin '/ rmax'<C ₄ (where C ₃ and C ₄ are predetermined constants), the red-eye portion is not circular and the contour line has no possibility of red-eye, and the red-eye recognition flag FLG is set to "0". , And returns from the subroutine (steps S36, 37, 39, 40). Furthermore, C ₃ <rmin '
When / rmax '<C ₄ , it is determined that there is a possibility of red eyes in the contour line, the red eye candidate flag FLG is set to "1", and the process returns from the subroutine (steps S37, 38, 40).

In the first embodiment described above, the red eye is automatically detected from the entire screen for the frame in which the red eye may occur. There is also. Therefore, a second embodiment which solves the problem of the first embodiment will be described in detail below.

FIG. 11 is a flow chart showing the operation of the second embodiment. When the red-eye correction system of the second embodiment starts operating, shading correction is first performed.
Then, here, the distortion due to the variation in the sensitivity of the sensor and the unevenness of the light amount of the light source is corrected (step S41).

Next, based on the image data detected by the color line sensor 6, the average transmission density (LATD) of the entire screen of the film 4 is measured (step S42). Then, from the LATD and the correction data from the LSI card 20, the density and color correction of the video data R, G, and B are performed to display the image of the film 4 on the monitor 14 (step S).
43-S45). Next, the corrected video data is stored in the video RAM 13 (step S46). Next, the magnetic data of the magnetic recording part 37 of the film 4 is read (step S47). Then, when there is a possibility of red eye occurrence in step S48, the monitor 14 displays
As described above, a display for instructing the setting of the red eye region is displayed at the lower left of the screen. When the operator traces the area around the eyes with a light pen, the correction portion area as shown in FIG. 7 is displayed as a contour line. Then, when the setting of the red-eye area is completed (step S4)
9) Next, the workstation 10 reads the area where the red eyes exist (step S50). Then, the subroutine "correction of red eyes" is executed from the designated area (step S51).

On the other hand, if it is determined in step S48 that there is no possibility of occurrence of red-eye, print density and R, G, B color corrections are performed (step S52). Then, after the reflectance-voltage linear signal (BGR) is converted into a density-voltage linear signal (YMC) by a γ correction unit (not shown), a gray component in the three color signals is separated (UCR) to generate a black signal. (Steps S53 to S55). Then, color correction masking is performed using a predetermined masking equation (step S5).
6). Next, after gradation correction is performed (step S5
7), edge enhancement (step S58), and sharpness correction such as smoothing is performed (step S59). Next, printing is executed based on the image data subjected to the above processing (step S59). Then, the film 4 is fed by one frame and the same processing is executed again (step S
60).

Although the case where the red-eye area is surrounded and designated is described above, the second embodiment is not limited to this, and the red-eye search area can be narrowed by only pointing one point near the red-eye. Greatly improves the processing speed.

Here, FIG. 12 is a diagram showing a basic configuration for arranging the transparent touch panel switch 54 on the screen of the monitor 14 and detecting the area designated by the operator at the workstation 10.

FIG. 13 is a diagram showing a detailed configuration of the transparent touch panel switch 54 and the monitor 14. In FIG. 13, the transparent electrode 56 and the lower electrode 58 are disposed below the graphic sheet 55, which is the operation surface of the transparent touch panel switch 54, via the spacer 57, and the upper electrode is provided at a position where the spacer 57 does not exist. An electrode X integrated with 56 and an electrode Y integrated with the lower electrode 58 are provided so as to face each other so that they do not normally contact each other, and one switch 54a of the transparent touch panel switch 54 is configured. Then, by pressing the graphic sheet 55 corresponding to the electrodes X and Y with a finger or a pen, the electrodes X and Y come into contact with each other and the switch is turned on. Further, the lower electrode 58 of the transparent panel switch 54 is adhered on the display screen of the monitor 14 with an adhesive tape 59.

Further, as shown in FIG. 14 (a), a plurality of electrodes X of the transparent panel switch 54 are provided on each line X ₁ ,
Connected X _2, X ₃ ..., the electrodes Y even also by plural Y
₁ , Y ₂ , Y ₃ ... Are connected and a plurality of switches are arranged in a matrix in which they cross each other. Now, in this transparent panel switch 54 arranged in a matrix, repetitive pulses as shown in FIG. 14B are sequentially transmitted from the workstation 10 to the lines X ₁ , X ₂ , and X ₃ via the IO port 9. Then, depending on which switch of the transparent panel switch 54 is turned on, which line pulse of the lines X ₁ , X ₂ , X ₃ ... Is output to which line of the lines Y ₁ , Y ₂ , Y ₃ ... It is decided whether it will be done. The device for designating the red-eye area may be designated by a well-known light pen input system or a mouse other than the transparent panel switch as described above.

As described in detail above, in the red-eye correction system of the present invention, the red-eye is automatically corrected only by roughly indicating the vicinity of the red-eye, so that the coordinates of the red-eye can be accurately indicated or the correction can be corrected. Red eyes are easily fixed without the need for advanced techniques.

[0043]

According to the present invention, it is possible to provide a red-eye correction system that converts an image containing red-eye into digital image data, automatically detects the red-eye, and restores the normal eye.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a red-eye correction system according to a first embodiment of the present invention.

FIG. 2 is an overall control block diagram of a camera.

FIG. 3 is a plan view of a film 36 having a magnetic recording portion 37.

FIG. 4 is a diagram for explaining a known contour line extraction method.

FIG. 5 is a diagram showing how coordinates are labeled on red-eye pixels.

FIG. 6 is a diagram illustrating a method of detecting an edge of an iris.

FIG. 7 is a diagram showing a display for instructing the setting of a red-eye area.

FIG. 8 is a flowchart showing the operation of the first embodiment.

FIG. 9 is a flowchart showing a sequence of a subroutine “red eye correction”.

FIG. 10 is a flowchart showing a sequence of a subroutine “red eye recognition”.

FIG. 11 is a diagram showing a configuration of a red-eye correction system according to a second embodiment of the present invention.

FIG. 12 is a diagram showing a basic configuration for arranging a transparent touch panel switch 54 on the screen of the monitor 14 and detecting an area designated by the operator at the workstation 10.

FIG. 13: Transparent touch panel switch 54 and monitor 14
It is a figure which shows the detailed structure with.

FIG. 14A is a time chart showing a plurality of electrodes of the transparent panel switch 54, and FIG. 14B is a time chart showing pulse output.

[Explanation of symbols]

1 ... Light source, 2 ... Diffusion plate, 3 ... Film carrier, 4 ... Film, 5 ... Lens, 6 ... Color line sensor, 7 ... Amplifying circuit, 8 ... AD converter, 9 ... I / O port, 10
... work station, 11 ... memory, 12 ... sensor scanning device, 13 ... video RAM, 14 ... monitor, 15 ... magnetic head, 16 ... signal processing circuit, 17 ... film feed roller, 18 ... notch detection sensor, 19 ... scanner Two
0 ... LSI card, 21 ... Printer.

Claims (3)

[Claims] 1. A scanner means for converting an image of a negative color film into digital image data, a red eye extracting means for extracting red eye from the digital image data, and a red eye correcting means for correcting the red eye extracted by the red eye extracting means. A red-eye correction system comprising: 2. A monitor means for visually inspecting an image of a negative color film before printing, a designating means for instructing a position in the vicinity of the red eye in a frame judged to be red eye as a result of the visual inspection by the monitor means. A red eye discriminating means for automatically discriminating a red eye from a nearby image designated by the designating means; and a red eye correcting means for correcting the red eye discriminated by the red eye discriminating means. Correction system. 3. A recording means for recording information on the presence or absence of the possibility of the occurrence of red eye on a negative color film, a reading means for reading the information recorded on the negative color film by the recording means, and before printing an image on the negative color film. Monitor means for displaying on the screen, an instruction means for manually instructing the position near the red eye to the frame where the red eye may occur, and an automatic red eye from the image in the vicinity instructed by the instructing means. A red-eye correcting system, comprising: a red-eye determining unit that determines the red-eye, and a red-eye correcting unit that corrects the red-eye determined by the red-eye determining unit.