JP2001024887A - Method for discriminating similar scene - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method by which a similar scene can be discriminated with accuracy. SOLUTION: A line CCD scanner 11 reads a DX code and picture data from a photographic film. A picture processor 12 discriminates a type of the film from the DX code and performs nonlinear correction on the picture data based on discriminated result. During a course of the nonlinear correction, the picture data are converted so that the data may become linear with respect to luminance of an object in a whole exposure area by correcting nonlinearity of the film. Then density histograms are found at every film picture from the corrected picture data and similarity is calculated through standardization and comparison of the histograms. Therefore, a similar scene can be discriminated appropriately without being affected by exposure conditions, such as under- exposure, over-exposure, at photographing time.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for determining a similar scene, and more particularly to reading a plurality of images recorded on a photographic material by image reading means and determining image processing conditions based on the read image data. The present invention relates to a similar scene determination method of an image processing apparatus.

[0002]

2. Description of the Related Art When a print is made from a photographic film using a lab device, a similar scene is determined in order to make prints of similar scenes in the same film uniform.
A method of correcting a parameter for adjusting the finish based on the similarity is generally known. As a similar determination method, a method of calculating an image feature amount from image data and finding a similar scene frame from the image feature amount (JP-A-54-26729, JP-A-56-153334).
Publication) and a method of comparing histograms.

[0003]

In the conventional method of judging similar scenes by comparing characteristic values, since the non-linearity of the photographic film is not taken into consideration, the characteristic curve of the photographic film has a characteristic curve corresponding to the foot or shoulder. In the case of a biased image, there is a problem that an appropriate determination cannot be made. For example, in a method of comparing similar scenes by comparing histograms, a case where a plurality of scenes photographed with proper exposure are compared with a case where a plurality of scenes photographed with underexposure are compared (a plurality of scenes identical to the proper exposure) is compared. , The result of the similarity determination will be different. This is because the gradation is destroyed in the case of under exposure. Further, when the same scene is photographed under different exposure conditions, it is difficult to determine the same scene by proper exposure, underexposure, and overexposure due to the characteristics of the photographic film.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and provides a similar scene judging method capable of properly judging whether a scene is a similar scene even with underexposed or overexposed frames. The purpose is to provide.

[0005]

According to a first aspect of the present invention, there is provided a method for judging a similar scene, comprising the steps of: performing a conversion process for correcting nonlinearity of a photographic material for each image; An image feature amount is calculated based on the image data after the correction processing, and a similarity between a plurality of images is calculated based on the image feature amount.

In the conversion processing for correcting the nonlinearity of the photographic material, it is preferable to use a conversion table stored for each type of the photographic material. The conversion table for correcting the non-linearity of the photographic material is cyan,
It is preferable to set independently for each color of magenta and yellow. Further, a common color may be used for each color. Preferably, the image feature amount is a density histogram.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS [Configuration of Digital Laboratory System]
FIG. 1 is a schematic configuration diagram of a digital laboratory system embodying the present invention. This lab system 10 is a line CCD
Scanner 11, image processing device 12, laser printer 1
3 and a processor 14. The line CCD scanner 11 and the image processing device 12 are integrated as an input device 16, and the laser printer 1
3 and the processor 14 are integrated as an output device 17.

The line CCD scanner 11 reads a film image recorded on a photographic film, such as a 135-type photographic film, a 110-type photographic film, an IX240-type photographic film, and a brownie-type photographic film. Can be read. This line CCD scanner 11
Reads the film image with a 3-line color CCD,
R, G, and B image data are output.

The image processing device 12 stores image data (scan data) output from the line CCD scanner 11.
Is entered. Also, a storage medium 18a such as a memory card
And an external device 18 such as another information processing device 18b,
Image data obtained by shooting with a digital camera,
Image data other than a film image, for example, obtained by reading a reflection original with a scanner, and image data generated by a computer (hereinafter, referred to as file image data) are input from outside.

The laser printer 13 includes R, G, and B laser light sources and a modulator. Then, the image processing apparatus 1
The laser beam is modulated in accordance with the recording image data input from step 2, the modulated laser beam is irradiated on the photographic paper, and the image is recorded on the photographic paper by scanning exposure. Further, the processor 14 performs each process of color development, bleach-fix, washing with water, and drying on the photographic paper after scanning exposure. As a result, an image is formed on the printing paper.

[Configuration of Image Processing Apparatus 12] Next, the configuration of the image processing apparatus 12 will be described. As shown in FIG. 2, the image processing apparatus 12 includes a line scanner correction unit 20A.
~ 20C. This line scanner correction unit 20
A to 20C are provided corresponding to R, G, and B data input from the line CCD scanner 11. The line scanner correction units 20A to 20C have the same configuration as each other, and are hereinafter simply referred to as the line scanner correction unit 20 without distinguishing them.

The line scanner correction unit 20 sequentially performs processes such as dark correction, density conversion, shading correction, and defective pixel correction. In the dark correction, the line CCD scanner 11
, The dark output level of the cell corresponding to each pixel is subtracted from the scan data from. In the density conversion, the data subjected to the dark correction is logarithmically converted into data representing the density of the photographic film 18. For shading correction, the photo film 18
The data on which the density conversion has been performed is corrected in pixel units according to the unevenness in the light amount of the light illuminating. In the defective pixel correction, among the data subjected to the shading correction, data of a cell (defective pixel) for which a signal corresponding to the amount of incident light is not output is interpolated from surrounding pixel data and newly generated.

The output of the line scanner correction unit 20 is selectively input to an input / output controller 22 and image processors 23A and 23B via a selector 21. Also,
The input terminal of the selector 21 is also connected to the data output terminal of the input / output controller 22.
2, the file image data input from the outside is input to the selector 21.

The optimum processing conditions for the image processing are as follows: whether the image data after the image processing is used for recording an image of photographic paper on the laser printer 13 or for displaying on a display means such as the display 43; Also changes depending on whether or not they are stored. For this purpose, the image processing apparatus 12 is provided with two image processors 23A and 23B. For example, in a case where the image data is used for recording an image on photographic paper and output to the outside, etc., a setup calculation unit 30A described later performs a setup calculation for each application and optimizes the application for each application. Are determined and output to the image processors 23A and 23B. Thereby, the image processor units 23A and 23B
In, different image processing is performed on the same fine scan image data.

The image processor 23A includes a memory controller 24, an image processor 25, and three frame memories 26A, 26B, 26C. Each of the frame memories 26A to 26C has a capacity capable of storing image data of one or more frame images. The image data input from the selector 21 is transferred to the three frame memories 26A by the memory controller 24.
To 26C. Note that the image processor unit 23B has the same configuration as the image processor 23A.

As shown in FIG. 3, the image processor 25 includes a first image processing unit 25A and an offset correction unit 25.
B, non-linear correction section 25C, image data normalization section 25D,
The second image processing unit 25E is sequentially connected. This image processor 25 includes a frame memory 26
The image data stored in the fine scan memory 26 (shown in FIG. 4) is fetched, and various image processing is performed by the auto setup engine 30 according to the processing conditions determined for each image.

The image processing executed by either the first image processing section 25A or the second image processing section 25E includes, for example, image enlargement / reduction, gradation conversion, density correction, color conversion, and hypertone processing. And image processing (standard image processing) for improving the image quality of an output image such as hyper sharpness processing. In the hypertone processing, the gradation of an ultra-low frequency luminance component of an image is compressed. In the hyper sharpness processing, sharpness is enhanced while suppressing graininess.

As shown in FIG. 2, the image processor 25
Are connected to the input / output controller 22. And
The image data subjected to the image processing is temporarily stored in the frame memory 26 and then output to the input / output controller 22 at a predetermined timing.

In the present embodiment, each film image is read twice by the line CCD scanner 11 at different resolutions. In the first reading at a relatively low resolution (hereinafter, referred to as pre-scan), even when the density of the film image is very low (for example, a negative image of an exposure under film in a negative film), the charge accumulated in the line CCD is reduced. The entire surface of the photographic film is read under the reading conditions determined so as not to cause saturation (the light amounts of the R, G, and B wavelengths of light applied to the photographic film, and the charge accumulation time of the line CCD). Data (pre-scan data) obtained by this pre-scan is supplied to the selector 21.
To the input / output controller 22.

An automatic setup engine 30 is connected to the input / output controller 22. The auto setup engine 30 includes a CPU 31, a RAM 32, and a ROM.
33, and an input / output port 34, which are connected to each other via a bus 35.

The pre-scan data output from the input / output controller 22 is stored in the pre-scan memory 3 shown in FIG.
After being temporarily stored in the memory 6, it is subjected to setup calculation processing in the auto setup engine 30 and simulation image display processing in the personal computer (personal computer) 40. As the pre-scan memory 36, the R of the auto setup engine 30 is actually used.
The AM 32 or the memory 42 of the personal computer 40 is used.

[Setup Operation Processing] FIG. 3 shows, among various functions realized by the CPU 31 of the auto setup engine 30, a function of performing a setup operation processing as a setup operation section 30A. The setup calculation unit 30A performs a setup calculation process as follows. First, the setup operation unit 30A
Determines the frame position of the film image based on the pre-scan data input from the input / output controller 22,
Data (pre-scan image data) corresponding to the film image recording area on the photographic film is extracted. Further, based on the pre-scanned image data, the size of the film image is determined, and the image feature amount such as the density is calculated to determine the reading conditions for performing high-resolution re-reading (fine scan).

The setup calculation section 30A performs image processing on image data (fine scan image data) obtained by performing fine scan by the line CCD scanner 11 based on pre-scan image data of a plurality of frames of film images. Are automatically determined by calculation, and the determined processing conditions are output to the image processor 25 of the image processor unit 23A.

In determining the conditions for the image processing, it is determined by the similar scene determination processing of the present invention whether or not there are a plurality of film images capturing similar scenes, and a plurality of film images capturing similar scenes are determined. If the image matches,
The processing conditions of the image processing for these film images are determined so as to be the same or similar.

Before the similar scene determination process, a conversion process for correcting the non-linearity of the photo film is performed. Due to this non-linearity correction process, even if there is an image biased on the foot or shoulder in the characteristic curve of the photo film,
The similarity determination can be appropriately performed.

[Nonlinearity Correction Processing] The ROM 33 of the auto setup engine 30 stores characteristic data FilmDa determined in accordance with the exposure-color density characteristics of the photographic film.
ta _j (d) is stored. Note that d represents the density and j
Represents any of R, G, and B. This characteristic data FilmData
_j (d) represents the conversion condition of the non-linear correction for converting the image data of the film image.
It has a conversion characteristic as shown in FIG. By this conversion, the density value represented by the image data of the film image stored in the photographic film is converted to the virtual value of the virtual film image obtained by exposing and recording the same subject as the film image on the virtual photographic film and performing development processing. The density value is set to be substantially equal to the density value represented by the image data. The virtual photo film at this time is the exposure amount for each of R, G, and B of the photo film.
(logH) -R, G, and R, respectively, equal to the linear portion of the color density (d) characteristic extended to the exposure region where the characteristic becomes non-linear.
5 is a virtual photographic film having the exposure amount-coloring density characteristic of B.

The characteristic data FilmData _j (d) can be determined, for example, as follows. As shown in FIG. 5 (A), for example, as shown in FIG. 5 (A), the exposure-color density characteristic of a negative film, which is a kind of a photographic film, is such that the color density changes in response to a change in exposure in a standard exposure region corresponding to normal exposure. Although linearly changing (directly proportional), in the exposure area corresponding to underexposure and the exposure area corresponding to overexposure, the gamma (change in color density with respect to change in exposure) increases as the distance from the standard exposure area increases. ) Gradually decreases, and the color density changes non-linearly with changes in the exposure amount. Therefore, as shown by the bold line in FIG. 5B, a virtual portion of the characteristic is extended to an exposure region where the characteristic becomes nonlinear with respect to the exposure-color density characteristic of the photographic film to be processed. A proper exposure-color density characteristic is set. This corresponds to the exposure-color density characteristic of the virtual photographic film described above.

Next, the exposure amount of the photographic film to be processed-
In the color density characteristic and the exposure-color density characteristic of the virtual photographic film, the color density corresponding to the same exposure is associated over the entire exposure area. For example, as shown in FIG. 5, the color density d1 of the photographic film to be processed and the color density d1 'of the virtual photographic film with respect to the exposure amount H1.
And the color density d2 of the processing target photographic film and the color density d of the virtual photographic film with respect to the exposure amount H2.
2 'is associated with the color density d3 of the photographic film to be processed and the color density d3' of the virtual photographic film with respect to the exposure amount H3.

Of the associated color densities, the color density of the photographic film to be processed is defined as the input density, the color density of the virtual photographic film is defined as the output density, and the density value of the film image recorded on the photographic film to be processed is calculated as A conversion condition for converting into a density value of a virtual image recorded on the virtual photo film is set. By performing the above processing for each of the R, G, and B component colors, characteristic data FilmData _j (d) representing a conversion condition as shown in FIG. 4A is obtained.

The exposure-color density characteristic of a photographic film is as follows.
In this embodiment, the above-described characteristic data FilmData _j (d) is set for each film type because the difference is in the unit of the film type of the photographic film. The ROM 33 stores characteristic data FilmData _j (d) of each film type in association with information for identifying the film type.

Note that the characteristic data storage section 30b stores FIG.
As shown in FIG. 3B, characteristic data FilmData _j (d) defined according to the exposure-color density characteristic representing the relationship between the exposure amount of the photographic film and the density of the reference color (for example, G) is stored. Is also good. In this case, the above-described non-linearity correction processing is commonly performed for each color using the exposure amount-coloring density characteristic.

[Similar Scene Judgment Processing] In the similar scene judgment processing, the similarity R between the film images is obtained, and the higher the similarity R is, the more the similarity is determined. As shown in FIGS. 6A and 6B, first, density histograms HEA and HEB are created for each of the film images EA and EB for which the similarity R is to be determined. The density histograms HEA and HEB are obtained from the image data on which the above-described nonlinearity correction processing has been performed. Then, as shown in FIG. 6C, the average densities CA and CB of the obtained density histograms HEA and HEB are obtained, these densities are adjusted, and the density histogram of the similarity determination target is standardized (CA, C
Histograms HEA, HE of each density so that B overlaps.
Shift B. Next, a frequency difference (hatched portion in FIG. 6D) at each density of the normalized density histogram is obtained, and based on this difference, a similarity R is calculated using the following equation.

S = T−Dnm (1) where S = 0 if S <0, and Dnm is the difference between the histograms of the n-th frame and the m-th frame (the hatched portion in FIG. 6D), T is an upper limit value of the similarity related value, and is a value set so that the similarity becomes “0” if Dnm is equal to or more than this value.

R = S / T (2) where R ranges from 0 to 1.0.

Based on the obtained similarity R, for example, when the similarity R is in the range of 0 ≦ R ≦ 1.0, it is determined as a similar frame. The same or similar image processing conditions are given to these similar frames, and the print finish of the frames determined to be similar is made uniform. Specifically, the similarity determination result is reflected in the image processing conditions as follows. When the parameter of a certain image processing condition is P, the parameter PN of the _Nth frame is calculated based on the parameters _PN-1 and _{PN + 1} of the preceding and succeeding N-1 frames and N + 1 frames by the following expression ( A similar frame correction process is performed according to 3).

[0036]

(Equation 1) Here, Ri is the similarity of the i-th frame, and Ri <1.0. Although the number of frames before and after the similar frame correction is arbitrary, 2-3 frames are preferably used.

[Calculation of Image Recording Parameters] The setup calculation section 30A shown in FIG. 3 calculates image recording parameters and outputs them at the same time as outputting recording image data to the laser printer 13. The image recording parameters define the gray balance when the laser printer 13 records an image on photographic paper, and are calculated based on the pre-scanned image data of the film image. The auto setup engine 30 similarly determines the processing conditions of the image processing by calculation for the file image data input from the outside.

As shown in FIG.
2 is connected to the laser printer 13 via the I / F circuit 52. When the image data after image processing is used for recording an image on photographic paper, the image processor unit 23
The image data subjected to the image processing in A is output from the input / output controller 22 to the laser printer 13 via the I / F circuit 52 as image data for recording. The auto setup engine 30 is connected to the personal computer 40. When the image data after the image processing is output to the outside as an image file, the image processing is performed in the image processor 23 and the image data is output to the input / output controller 22.
Is output to the personal computer 40 via the auto setup engine 30.

The personal computer 40 has a CPU 41 and a memory 4
2, display 43, keyboard 44, hard disk 45, CD-ROM driver 46, transport control unit 47,
An expansion slot 48 and an image compression / decompression unit 49 are provided, and these are connected to each other via a bus 50.

FIG. 3 shows, among the various functions realized by the CPU 41 of the personal computer 40, a function relating to the simulation image display processing by a plurality of blocks (that is, a prescan image data processing section 40A, an image display section 40B, and a key) for each function. The correction input section 40C) is shown separately. The pre-scan image data processing unit 40A fetches the pre-scan image data from the pre-scan memory 36 and fetches the processing conditions of the image processing determined by the setup calculation unit 30A. The prescan image data is data extracted from the prescan data by the setup calculation unit 30A and stored again in the prescan memory. Then, the pre-scan image data processing unit 40A targets the fine processor image data to the image processor 25 based on the fetched processing conditions.
The image processing equivalent to the image processing performed in (1) is performed on the prescanned image data to generate simulation image data.

The image display section 40B includes a display 43 (see FIG. 2), and displays the simulation image data generated by the prescanned image data processing section 40A on the display 43. The operator observes the simulation image and performs a test for image quality and the like. Then, the operator inputs information for instructing the correction of the processing condition from the keyboard 44 as the test result.
The key correction input unit 40 outputs the input correction instruction information to the auto setup engine 30. This allows
The setup calculation unit 30A performs recalculation of image processing conditions and the like.

On the other hand, the transport controller 47 shown in FIG. 2 is connected to the film carrier 20 and controls the transport of the photographic film by the film carrier. The film carrier 20 is set on the line CCD scanner 11, and transports the photographic film in the longitudinal direction.

A driver for reading / writing data from / to an information storage medium such as a memory card, and a communication control device for communicating with other information processing equipment are connected to the personal computer 40 via the expansion slot 48. You. When image data for output is input from the input / output controller 22 to the outside, the image data is output to the outside as an image file via the expansion slot 48. When file image data is input from outside via the expansion slot 48, the input file image data is output to the input / output controller 22 via the auto setup engine 30. In this case, the input / output controller 22 converts the input file image data into the selector 21.
Output to

Next, as an operation of the present embodiment, a case where a film image recorded on a photographic film is read by the line CCD scanner 11, and various image processes are performed on the image data obtained by the reading and output will be described. I do.

As shown in FIG. 2, first, the line CCD scanner 11 reads a film image recorded on a photographic film twice: prescan and fine scan. In the prescan, reading is performed on the entire surface of the photographic film. When the pre-scan data is input to the image processing device 12, the line scanner corrector 20 performs various processes of dark correction, density conversion, shading correction, and defective pixel correction.

The pre-scan data output from the line scanner correction unit 20 is temporarily stored in a pre-scan memory 36 (see FIG. 3) via the selector 21 and then taken in by the auto setup engine 30. Setup operation processing is executed. Hereinafter, this setup calculation process will be described with reference to FIG.
This will be described with reference to flowcharts of FIGS.

In step 100, from the pre-scan data fetched from the pre-scan memory 36, data in the area where the DX code barcode is recorded on the photographic film to be processed is extracted. In the next step 102, the contents of the DX code recorded on the photographic film to be processed are determined from the data extracted in step 100, and the film type of the photographic film to be processed is determined based on this content. In step 104, ROM 33 from the (characteristic data storage unit 30B) characteristic data for each film type stored in FilmData _j (d), film species corresponding characteristic data FilmData _j of the photographic film to be processed (d )
Take in. Then, in the next step 106, the acquired characteristic data FilmData _j (d) is set as a conversion condition for nonlinearity correction.

In step 108, data of an unexposed area (plain area) on the photographic film to be processed is extracted from the pre-scan data fetched from the pre-scan memory 36, and R, G, B represented by the extracted data Of the density value of the photographic film to be processed
Set as eDns _j .

In the next step 110, the previous step 1
The minimum densities of R, G, and B defined by the characteristic data FilmData _j (d) captured in step 04 are set as film-based densities Basedns _j . And step 1
In 12, the reference value BaseData _j of the film base density BaseDns _j and film base density, is stored as the offset correction value for the photographic film to be processed.

In the following steps 114 to 120, processing is performed for each film image. That is, in step 114, the recording position (frame position) of the one-frame film image on the photographic film is determined based on the pre-scan data fetched from the pre-scan memory 36,
Film image data corresponding to the film image recording position from the pre-scan data based on the determined frame position IMA
GE _j (x) is cut out (x represents an ID for identifying each pixel). This film image data IMAGE _j
(X) (pre-scan image data) corresponds to the read image data according to the present invention.

In step 116, the disconnection is made in step 114.
IMAGE image data _j (X)
Using the offset correction value stored in step 112 of
The data of each pixel is defined as a unit according to the following equation (3).
To perform offset correction processing.

IMAGE2 _j (x) = IMAGE _j (x) −BaseDns _j + BaseData _j (3) The density of the film base of the photographic film varies depending on the development conditions when the photographic film is subjected to the development processing. ,
This change in film-based density is based on the film image data.
It is added to IMAGE _j (x) as an offset. Further, even when the amount of light emitted from the light source used for reading is different from the expected amount of light when reading a photographic film by the line CCD, the difference in the amount of light is reflected by the film image data IMA.
Add as an offset to GE _j (x). In contrast, in the above offset correction processing, the film base density Bas
The film image data IMAGE _j (x) is corrected according to the difference between eDns _j and the reference value BaseData _j of the film base density, so that the development condition of the development processing for the photographic film to be processed is changed and the photographic film is read. IMAGE _j (x)
Can be corrected.

In the next step 118, offset correction
Film image data IMAGE2 after processing _j (X)
Characteristic data FilmData captured in step 104 of_j 
Using (d), the data of each pixel is calculated according to the following equation (4).
Non-linearity correction processing is performed for each of R, G, and B in data units.
U.

IMAGE3 _j (x) = FilmData _j (d) + (IMAGE2 _j (x)) (4) By the above-described nonlinear correction processing, the nonlinearity of the exposure amount-coloring density characteristic of the photographic film to be processed The effect due to is eliminated. Then, the film image data IMAGE3 _j (x) after the offset correction processing is converted into image data (film image data IMAGE3 _j (x)) substantially equal to the image data of the virtual film image recorded on the virtual photograph film. Become. The virtual photographic film has an exposure-color density characteristic (see a thick line in FIG. 6B) in which the color density changes linearly with the change in the exposure over the entire density range.

In the next step 120, it is determined whether or not the above-described processing in steps 114 to 118 has been performed for all the film images recorded on the photographic film to be processed. If the determination is negative, the process returns to step 114, and steps 114 to 120 are repeated until the determination in step 120 is affirmed. Thereby, the film image data IMAGE3 _j (x) after the non-linear correction processing is obtained for all the film images.

The exposure-color density characteristics of the photographic film are different for each of R, G, and B even when the photographic film is developed under the standard developing conditions. Deviates from the standard development conditions, the actual exposure-color density characteristics of the photographic film further vary for each of R, G, and B. Therefore, the color balance of the image represented by the film image data IMAGE3 _j (x) after the non-linearity correction processing is determined by the difference between the R, G, and B of the actual exposure-color density characteristics of the photographic film. At the time of exposure (at the time of exposure recording of the subject image on the photographic film) (the gray balance is lost). Therefore, in the next step 122 and subsequent steps, the image data is normalized so that the color balance of the image represented by the image data substantially matches the color balance of the subject image.

In step 122, each film image is analyzed based on the film image data after the non-linearity correction processing of each film image. As a result, from the film image data of each film image, a pixel (gray candidate) corresponding to a portion where the hue is estimated to be an achromatic color (gray or the like) on the subject image captured and recorded on the processing target photographic film as the processing target film image Points) are extracted. The extraction of the gray candidate points is performed, for example, as follows.

Based on the film image data of each film image, the high density side R, G, B reference density representing the high density side achromatic color condition and the achromatic color condition on the low density side are defined for each film image. The low-density R, G, and B reference densities to be represented are respectively set. As these reference densities, for example, the color of the pixel having the maximum three-color average density in the film image and the hue of the pixel having the minimum three-color average density (the hue of the portion corresponding to the pixel on the subject image) are regarded as achromatic. Regarding, R of the pixel,
The G and B density values can be used as the R, G and B reference densities on the high and low density sides. Further, in the density histogram of the three-color average density, the tint of a pixel whose cumulative frequency from the maximum value and the minimum value of the three-color average density is a predetermined value is regarded as an achromatic color, and the R, G, and B density values of the pixel are determined as described above. The reference density may be used. Further, the average density for each of R, G, and B of a plurality of pixels whose cumulative frequency from the maximum value of the three-color average densities is within a predetermined range is defined as the high-density R, G, and B reference densities. The average density for each of R, G, and B of a plurality of pixels whose cumulative frequency from the value is within a predetermined range may be used as the low-density R, G, and B reference densities.

Next, in order to improve the accuracy of the R, G, and B reference densities, the average values of the R, G, and B reference densities for each of R, G, and B set for each film image are set. (High-density-side average R, G, B reference density), low-density-side R, G, and B reference density average values for each of R, G, and B (low-density-side average R, G,
B reference density).

Subsequently, based on the high-density-side average R, G, and B reference densities and the low-density-side average R, G, and B reference densities, the three-color average density of each pixel of the film image and the tint of each pixel are determined. The achromatic condition representing the relationship with the R, G, and B densities when a is an achromatic color is determined. This is, for example, as shown in FIG. 10, the high-density-side average is displayed on the color coordinates (in FIG. 10, the color coordinates with the R density-G density on the horizontal axis and the G density-B density on the vertical axis). R, G, B reference densities and low-density average R, G, B reference densities are plotted, for example, both are connected by a straight line (called an achromatic line), and R is selected from the film image data of all the film images. , G, B densities are extracted from a large number of pixels that fall within a certain range centered on the achromatic line (for example, a range surrounded by a broken line in FIG. 10), and three-color average densities are extracted from the extracted pixel data. And the achromatic condition representing the relationship between R, G, and B density of each pixel can be determined by the least square method, regression analysis, or the like.

The achromatic condition is that an achromatic pixel in the film image (a pixel corresponding to a gray portion in the subject image)
Represents the density balance of R, G, and B (color balance in each density range), and extracts a pixel matching the achromatic color condition as a gray candidate point. The method for extracting gray candidate points is not limited to the above, and various known extraction methods can be applied. In step 122, from among the gray candidate points extracted as described above,
A pixel in a density region corresponding to a linear portion in the exposure amount-coloring density characteristic of the photographic film to be processed is extracted as a final gray candidate point.

In the next step 124, the weighting factors α1 and α2 used in the normalization processing of the image data are determined. The weight coefficient α1 is for the normalization condition obtained from the film image data, and the weight coefficient α2 is for the normalization condition obtained from the characteristic data. This weight coefficient α
1, α2 can be determined so that the value of the weight coefficient α1 increases as the number of gray candidate points increases.

In order to match the color balance of the image represented by the image data with the color balance of the subject image, the exposure amount of each color (R, G, B) at the time of shooting and each color on the image represented by the image data This can be realized by converting image data in units of data of each pixel (normalization process) so that the relationship with the density of (R, G, B) matches (overlaps) in each color.

In the present embodiment, since the normalization process is performed on the film image data after the non-linearity is corrected, the above relations for each of the R, G, and B colors are linear. Therefore, the normalization processing according to the present embodiment can be realized by performing linear conversion on the film image data in units of R, G, and B color data. In the following steps 126 and 128, a normalization condition for normalizing the image data by linear conversion is determined based on the data of the gray candidate points.

First, in step 126, a first normalization condition for normalizing so that the average densities Dr, Dg, and Db of the R, G, and B gray candidate points are all equal (Dr = Dg = Db) is set. R, based on a specific component color (for example, G)
It is obtained by calculation for each of G and B. The first normalization condition includes a gain Ganma1 _j and an offset amount Offset1 _j , which are parameters of a linear conversion equation for performing normalization.

In the case where the number of the gray candidate points is small, for example, the average densities Dr, D
In some cases, the difference between g and Db does not accurately represent the deviation between the R, G, and B colors in the relationship between the exposure amount at the time of shooting and the density on the image represented by the film image data. In this case, even if the normalization process is performed using only the first normalization condition, the color balance cannot be accurately corrected.

Therefore, using the second normalization condition including the gain Ganma2 _j and the offset amount Offset2 _j which are the parameters of the linear conversion formula, the conversion curves of the conversion conditions for each color are almost matched (overlap). I do. The second normalization condition is set in advance and stored in the characteristic data storage unit 30B in advance.

In the next step 130, the second normalization condition corresponding to the film type of the photographic film to be processed is fetched. In the next step 132, the parameters of the normalization condition calculated in step 128 and the parameters of the normalization condition fetched in step 130 are stored in the following (5),
The parameters are calculated by weighting and adding according to equation (6). Ganma _j = α1 × Ganma1 _j + α2 × Ganma2 _j ... (5) Offset _j = α1 × Offset1 _j + α2 × Offset2 _j ... (6)

In step 134, the film image data
IMAGE3 _j (x) is fetched, and the R, G, and B standardization processing is performed using the parameters obtained in (5) and (6) above and the following equation (7). Do.

IMAGE4 _j (x) = Ganma _j × IMAGE3 _j (x) + Offset _j (7) By the above-described normalization processing, an image to be processed resulting from a difference in the original exposure amount-coloring density characteristic of the photographic film Of the color balance (balance of density for each of R, G, and B) can be corrected. As a result, image data is obtained in which the relationship between the exposure amount of each color of R, G, and B at the time of imaging and the density on the image represented by the image data is linear and almost coincident with each other. Therefore, it is possible to obtain film image data IMAGE4 _j (x) that accurately represents a subject image at the time of shooting.

In the above-mentioned normalization processing, the parameters (Ganma1 _j , Offset1 _j ) of the first normalization condition obtained from the film image data are based on the number of gray candidate points finally extracted from each film image. , Parameters of the second normalization condition (Ganma2 _j , Offset2
_j ) and the parameters of the normalization condition (Ganma _j ,
Offset _j ) is determined, so even if the number of gray candidate points extracted from each film image is small and the correction accuracy by the normalization process is estimated to be low, the weight for the parameter of the first normalization condition is By reducing the size, the color balance of the image represented by the image data can be corrected with high accuracy.

In step 136, a density correction value of the film image is calculated based on the film image data after the normalization processing. This density correction value is obtained by using the following equation (8) and a known density correction calculation equation that is set in advance. DHos _j = f (IMAGE4 _j (x)) (8)

In the next step 138, it is determined whether or not the above-described processing in steps 122 to 136 has been performed on all the film images recorded on the photographic film to be processed. If the determination is negative, the process returns to step 122, and steps 122 to 136 are repeated until the determination in step 138 is affirmative. Thus, the normalization condition parameters Ganma _j , Offs
et _j and the density correction value DHos _j are obtained respectively.

In step 140, gray tone balance correction processing is performed on similar frames. In this correction processing, the parameters calculated in step 132 are subjected to weighting processing according to the similarity using the following equations (9) and (10).

(Equation 2)

(Equation 3)

In step 142, a density correction process is performed on similar frames. This correction processing is performed in step 13
The density correction value calculated in 6 is subjected to weighting processing according to the similarity using the following equation (11).

(Equation 4)

In step 143, processing conditions for image processing on the fine scan image data are determined. The determined processing conditions of the image processing are stored in the RAM 148 in association with the frame number information for identifying the specific film image. In step 144, the film image data IMAGE5 _j (x) is calculated based on the normalized condition parameters and the density correction values subjected to the similar frame correction. In step 145, the reading condition of the line CCD scanner at the time of the fine scan is calculated based on the image feature amount.

In step 146, it is determined whether or not the processing in steps 140 to 144 has been performed on all the film images recorded on the photographic film to be processed.
If the determination is negative, the process returns to step 140, and the processes of steps 140 to 144 are repeated thereafter. If the determination in step 146 is affirmative, the process proceeds to step 148,
Film image data IM after normalization processing of each film image
AGE5 _j (x) is transferred to the personal computer 40 via the pre-scan memory 36, and the image processing conditions for the fine scan image data of each film image are transferred to the personal computer 40, thus ending the setup calculation process.

Film image data IMAG of each film image
When the processing conditions of E5 _j (x) and the image processing are transferred, the personal computer 40 performs image verification processing. That is, as shown in FIG. 3, the prescanned image data processing unit 40A performs image processing performed by the image processor 25 on the fine scan image data based on the processing conditions of the image processing transferred from the auto setup engine 30. Is performed on the film image data to generate simulation image data. Also,
In the image processing device 12, based on the simulated image data generated by the prescanned image data processing unit 40A, the image processing device 12 represents a finish when a print is created using fine scan image data after image processing has been performed by the image processor 25. The simulation image is displayed on the image display unit 40B (display 43).

The operator observes the simulation image and checks whether the frame position is proper or not, and whether the image quality of the simulation image is proper or not. This determination result is input via the key correction input unit 40C (keyboard 44).

When an instruction to correct the frame position or the image processing conditions is given to a specific film image, the auto setup engine 30 executes the input operation of the setup calculation process described above for the specific film image. The processing corresponding to the corrected instruction is performed again, and the film image data reflecting the corrected instruction and the processing conditions of the recalculated image processing are transferred. As a result, the corrected simulation image is displayed again on the display 43.

This redisplayed image is observed, and if a correction is necessary again, a similar correction operation is performed. If the correction is appropriate, information indicating the end of the test is input, and the image test process is terminated.

When the pre-scan for the photographic film is completed, the fine scan is performed by the line CCD scanner 11. For this fine scan,
The reading conditions for each film image are notified from the auto setup engine 30 to the line CCD scanner 11. Therefore, each film image is fine-scanned according to the reading conditions.

The processing conditions for image processing for each film image are notified from the auto setup engine 30 to the image processor 25 when the fine scan image data of each film image is input from the line CCD scanner 11. The image processor 25 performs image processing of the notified contents on the input fine scan image data of each film image.

The offset calculation unit 25B of the image processor 25 uses the offset correction value stored in the above-described step 112 of the setup calculation processing to obtain (3)
The offset correction processing is performed according to the equation. Further, the non-linearity correction unit 25C performs step 10 of the setup operation.
Characteristic data FilmData _j (d) set as conversion conditions in 6.
And the nonlinearity correction processing is performed according to the equation (4). Further, the image data normalizing unit 25D determines the parameter (Gan) calculated in step 132 of the setup calculation process.
ma _j , Offset _j ), and performs a normalization process according to equation (4). Through these processes, the fine scan image data is converted into image data that accurately represents the subject image at the time of shooting.

The fine scan image data input to the image processor 25 is subjected to various types of image processing by the first image processing unit 25A and the second image processing unit 25E in addition to the above-described image processing. The image data is output from the image processor 25 as output image data. The second image processing unit 25E also performs a density correction process based on the density correction value of each film image obtained according to the similarity in step 142. The output image data is stored in the laser printer 13.
Is used to record an image on photographic paper, is used to display an image on the display 43, or is stored in an information recording medium such as a memory card via the expansion slot 48.

In the above embodiment, the similarity determination processing is performed based on the pre-scan image data, but instead, the similarity determination processing may be performed based on the fine scan image data.

In the above-described embodiment, the difference in frequency between the densities of the density histogram is obtained, and the similarity is obtained based on the difference. Alternatively, the similarity of the density histogram may be obtained by using a well-known pattern recognition technique. It may be determined whether or not it is performed. Alternatively, it may be determined whether or not the scenes are the same by using the same scene frame determination method described in Japanese Patent Application Laid-Open No. 54-27729. Also in this case, after correcting the nonlinearity of the photographic film and converting the image data so as to be linear with respect to the luminance of the subject, it is determined whether or not the scenes are the same. Further, before the similarity determination in the exposure control method described in JP-A-56-153334, the above-described nonlinearity correction may be performed to implement the present invention.

In the above embodiment, the present invention is applied to a digital print system. In addition, the present invention is applied to an analog print system in which transmitted light of a photographic film is imaged on a photosensitive material by a printing lens and printed. May be. In this case, the obtained similarity determination result is reflected in the exposure control amount so that a similar finish can be obtained for similar scenes. For example, one of the plurality of images is changed so that the image processing condition between the plurality of images approaches based on the obtained similarity. In addition,
Changing one of the image processing conditions of the plurality of images based on the similarity so that the image processing conditions of the plurality of images approach each other is performed by a digital print method in addition to the analog print method. You may.

In the above embodiment, the DX code is determined based on the pre-scan data. In addition, the DX code is read from the photographic film using a bar code reader, and the film type is determined based on the DX code to correct the nonlinearity. May be performed.

[0090]

According to the present invention, for each image, the non-linearity of the photographic light-sensitive material is corrected, and the image data is converted so that the brightness of the subject becomes linear in the entire exposure range, and then the similar scene is obtained. Is determined, without being affected by the exposure conditions at the time of shooting, such as underexposure and overexposure, etc.
Similar scenes can be properly determined. Therefore, even when the same scene is photographed under different exposure conditions, it is possible to properly determine whether or not the scene is the same. In addition, the print quality is stabilized by making the print finish uniform for frames determined to be similar. In addition, the image feature amount is defined as a density histogram, and the similarity is determined based on a difference between the histogram and the pattern recognition.
It can be determined with high accuracy whether or not they are similar.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a digital laboratory system embodying the present invention.

FIG. 2 is a block diagram illustrating a schematic configuration of an image processing apparatus of the digital laboratory system.

FIG. 3 is an auto setup engine of the image processing apparatus,
FIG. 2 is a block diagram showing functions of a personal computer divided into blocks and showing an internal configuration of an image processor.

FIG. 4A is a diagram illustrating an example of a conversion characteristic represented by characteristic data according to the first embodiment, and FIG. 4B is a diagram illustrating an example of a conversion characteristic represented by characteristic data according to the second embodiment;

5A is a diagram illustrating an example of an exposure amount-coloring density characteristic of a photographic film to be processed, and FIG. 5B is a diagram illustrating an example of an exposure amount-coloring density characteristic of a virtual photographic film.

FIG. 6 is an explanatory diagram of a process of determining a similarity using a density histogram.

FIG. 7 is a flowchart illustrating a setup calculation process.

FIG. 8 is a flowchart illustrating a setup calculation process.

FIG. 9 is a flowchart illustrating a setup calculation process.

FIG. 10 is a diagram showing an example of an achromatic condition for extracting a gray candidate point.

[Explanation of symbols]

 Reference Signs List 10 digital laboratory system 11 line CCD scanner 12 image processing device 25 image processor 30 auto setup engine 32 input / output controller 40 personal computer

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B057 BA02 CA01 CA08 CA12 CA16 CB01 CB08 CB12 CB16 CC01 CE11 CE17 CH01 CH07 DA12 DB02 DB06 DB09 DC23 DC34 5C077 LL19 MM03 MP08 PP15 PP35 PP37 PQ12 PQ18 PQ19 PQ23 TT02

Claims (4)

[Claims] An image reading unit reads a plurality of images recorded on a photographic photosensitive material, and determines whether or not the plurality of images are similar based on image data of the read image. In the determination method, a conversion process for correcting the nonlinearity of the photographic light-sensitive material is performed on each of the images, an image feature amount is calculated based on the image data after the nonlinearity correction conversion process, and a plurality of image feature amounts are calculated based on the image feature amount. A method for determining a similar scene, comprising calculating a degree of similarity between images. 2. The similar scene judging method according to claim 1, wherein the conversion processing for correcting the nonlinearity of the photosensitive material uses a conversion table stored for each type of the photosensitive material. 3. The conversion table for correcting non-linearity of the photographic light-sensitive material is independent or common for each color of cyan, magenta and yellow.
A method for determining the described similar scene. 4. The method according to claim 1, wherein the image feature amount is a density histogram.