US 4,769,695 AGrant

Method of Detecting Principal Subject Images and Determining Exposures in Color Printing

Issue Date:1988-09-06

•16 Claims

•8 Drawing Sheets

Abstract

Three color densities of a color original are measured for a number of picture elements into which the color original is divided. A part of the color original that includes at least one picture element is selected, with whose color based on the measured densities thereof the color of each picture element based on the measured densities is compared to detect picture elements having colors identical or similar to the color of the selected part of the color original. The picture elements of the identical or similar colors can define an extent of image which determines a principal subject image and the remainder of which is background. For the individual image, at least one characteristic value peculiar to the image is then calculated by utilizing the measured color densities of picture elements, and on the basis of the characteristics values the scene of the original is classified to select one of a plurality of previously provided exposure calculating formulas in order to determine the proper exposure for obtaining color prints of proper color balance and density.

Metadata

Assignee

Fuji Photo Film Co., Ltd.

Inventor

Takaaki Terashita

Application Information

Application Number:US 8554129

Filing Date:1986-04-24

Priority Date:1986-02-17

Art Unit:262

Classifications

IPC:

H04N 146G03B 2780

Field of Search:

35535835;3875;76;77;78;80

Patent Drawings (8 sheets)

Description

Background of the Invention

The present invention relates to a method of detecting a principle subject image and determining exposures in color printing, and more particularly to a method in which the extent of a principle subject image of a color original is detected based on color.

In making positive color prints on color paper from color originals such as color negatives, color positives or the like, it is usual to control the exposure or the proportion of the primary color (blue, green and red) components of printing light in accordance with the large area transmittance densities (hereinafter referred to as LATDs) of the primary colors of the color original from which the color prints are made. This printing system is well known in the art as an LATD printing system.

It is also well known in the art that almost all ordinary color originals include a background besides a principal subject image.

Since in the above-described conventional LATD printing system, the determination of exposures is independent of the primary color densities of the principal subject image of color originals, the obtained prints are not always satisfactory because of improper color balance and/or color density of the principal subject image thereof.

In order to solve the above-described problem in the LATD printing system, there has been proposed a method of color printing in which the image of a color original is displayed on a CRT screen of a color monitor and a point on a principal subject image such as the human face is specified by means of a light pen in order to measure the primary color densities of the point according to which the proportion of the primary color components of printing light are controlled. This method is disclosed in, for example, Japanese Patent Unexamined Publ. (Kokai) No. 58-9136.

In this method, however, there is a problem that, since only a part of the principal subject image of the color originals is specified to be measured, the color densities cannot be obtained accurately for the whole area of the principal subject image. This problem can be solved by specifying the boundary between the principle subject image and its background image by using a position specifying means such as a light pen. This solution is, however, undesirable and hardly practicable owing to a troublesome and time-consuming operation of specifying the boundary. As a matter of course, in order to determine the exposure or the proportion of the primary color components of printing light based on the color densities of a part of the principal subject image of color originals, it is necessary to obtain exact primary color densities. However, this is very difficult in practice because there are a large number of color originals of high contrasts and/or small pricipal subject images. In addition, the fact of disregarding the background when determining the exposure results in reproducing unfaithful prints.

There has also been disclosed in, for example Japanese Patent Publ. (Koho) No. 55-29412, a method of color printing for solving the above-described problem in which a color original is divided into plural image sections, one of them being chosen as a principal subject image. The chosen image section is scanned at a number of points (picture elements) to obtain a plurality of color densities thereof, from which at least one kind of characteristic density value is calculated. Based on the characteristic density values thus obtained and the LATDs of the color original, the exposure or the primary color components of printing light is controlled.

In this method, it is also necessary to obtain exact characteristic density values. Nevertheless, because of the relatively extensive divided sections, there is often included in the chosen section a part of a background image besides the pricipal subject image as a result of a visual analysis of an operator. For example, when the principal subject matter of the human figure has been photographed with a back light, the chosen section of the color original sometimes contains therein parts of the human figure and sky. As a result, confusion arises as to which is the intended principal subject matter: the human figure or the sky, and as a result, wrong characteristic density values will be used to control the exposure in color printing. In addition, the use of only the characteristic density values makes it impossible either to recognize color originals whose scenes are photographed with a back light or to calculate the value of exposure to be corrected in accordance with the LATDs.

There has been disclosed in, for example, Japanese Patent Unexamined Publ. (Kokai) No. 59-164547 still another method of color printing in which, in order to calculate exposures, one of a plurality of different formulas previously provided is selected through a selection key according to differences in density between the principal subject and its background images of color originals, the difference being judged by an experienced visual observer. In this method, it is possible to control the exposure with a high accuracy depending on closely grouped scenes in which the color originals are classified. This classification of color original, which is effected though visual analysis by the operator, is attended by individual differences in visual analysis, uncertain judgements and the like, and the provision of the closely grouped scenes is not always efficient. Consequently, the method is, although basically highly efficient, useful with only several scenes.

There has also been disclosed in, for example, Japanese Patent Unexamined Publ. (Kokai) Nos. 52-156624 and 53-145620 a method of color printing in which a particular extent of color on negatives is defined corresponding to flesh color on positives (this particular extent of color is hereinafter referred to as flesh color). A color original is scanned to obtain the primary color densities for each of a number of points (picture elements) on the color original in order to automatically judge whether or not each picture element has its color falling within the limits of flesh color previously defined, so as to control the exposure based on the mean color densities of the picture elements of flesh color. According to this method, in case the color originals include walls or the like which have a color similar to the flesh color of the human body, it is apt to take the walls for a part of the human body and to confuse gray color with flesh color, this color confusion being attributable to changes on standing in the characteristics of the film and the illumination lamps, resulting quite often in incorrect exposure controls in color printing.

Summary of the Invention

Accordingly, the primary object of the present invention is to provide a method for automatically determining the precise extent of a principal subject image of color originals.

Another object of the present invention is to provide a method for determining exposures in color printing so as to make exposure controls both with high accuracy and with ease.

In accordance with the present invention a color original to be printed is measured for its red, green and blue densities at a large number of points on (picture elements of) the color original. In order to determine the extent of a principal subject image of the color original, the measured densities, or the color obtained from the measured densities, of each picture element are comapared with the corresponding color densities, or the color obtained from the color densities, of a part of the color original which is specified by use of an image specifying means to judge whether or not the respective picture elements have colors identical or similar to the color of the specified part of the color original. From these color judgements, a principal subject image of the color original is determined to have an extent which is defined by the picture elements with colors identical and similar to the color of the specified part.

According to an important feature of the present invention, the extent or range of a particular principal color or colors of a color original is first defined. When the specified part is judged to have a color falling within the extent or range of the particular color in question, each picture element of the color original is judged as to whether or not its color falls within that range.

It should be noted that in this specification, the term "color density" means the density of monochromatic color such as red, green or blue and is obtained in such a way that the light either transmitted by or reflected from color originals is measured, after passing through a filter which permits a narrow band of wavelengths of light to pass therethrough, by using an image sensor whose outputs are processed to provide density values, and that the term "color" means what is obtained according to the combination of three primary color densities but is independent of brightness (luminous density). Therefore, colors whose three primary color densities are correspondingly identical to each other or whose ratios of the three primary color densities are identical to each other are defined as being identical to each other; colors whose three primary color densities are correspondingly close to each other or whose ratios of three primary color densities are close to each other are defined to be similar to each other.

According to a method of determining printing exposures of the present invention a color original to be printed is divided into a number of picture elements of which each one is measured as to its red, green and blue densities. In order to determine the exposure for the color original, the measured color densities, or the color obtained from the same, of each picture element are compared with the corresponding color densities, or the color obtained from the same, of a particularly specified part of the color original to judge whether the respective picture elements have colors identical or similar to the color of the specified part of the color original. Based on this judgement the extent of the image is defined by the picture elements both with the identical color and with similar colors and determined to be a principal subject image. Either the whole or a part of the remaining image is determined to be background image. At least one kind of image characteristic value is extracted from the principal subject image and from the background image, based on the color densities of the picture elements, in order to classify the scene of the color original. Based on the classified scene one adequate formula for calculating the printing exposure is selected from previously provided various formulas corresponding to scenes into which color originals are to be classified.

According to an important feature of the present invention, the extent of a principal subject image can be automatically determined by specifying only a part of the image of a color original with particular colors, which enables printing operations to be easily performed. Furthermore, the extraction of the image characteristic values from the principal subject and background images of color originals, on the basis of which the color originals are classified into scenes and the appropriate formula is chosen from among formulas previously provided corresponding to the scenes in order to calculate the exposures, contributes to avoiding individual differences between operators in visual analysis, and so makes possible accurate exposure controls in color printing. Since the same formula is applied to color originals which have relationships between their principal subject and background images similar to each other, the selection of a formula is highly accurate.

Brief Description of the Drawings

FIG. 1 is a schematic view of a color printer to which the present invention is applied;

FIG. 2 is an explanatory illustration of an example of an image sensor;

FIG. 3 is a functional block diagram illustrating an example of a microcomputer used in the color printer of FIG. 1;

FIG. 4 shows color maps which define flesh color and gray color in a two-dimensional coordinate system the axes of which respectively represent the density differences between green and blue densities, and between red and green densities;

FIG. 5 is a color map which defines the extent of similar color in a two-dimensional coordinate system the axes of which respectively represent the density differences between green and blue densities, and between red and green densities;

FIG. 6 is a flow chart showing a sequence of exposure control in the color printer of FIG. 1;

FIG. 7 is a explanatory illustration of an example of an image extent table;

FIG. 8 is an explanatory illustration of an example of a touch panel sensor;

FIG. 9 is a flow chart showing a sequence of judging position and color of a standard picture element;

FIG. 10 is a flow chart showing a sequence of determining the extent of a principal subject image;

FIG. 11 is a flow chart showing a sequence of calculating image characteristic values and scene classification information; and

FIG. 12 is a flow chart showing a sequence of scene classification and the selection of exposure calculation formulas.

Detailed Description of the Invention

Referring now to FIG. 1 showing a color printer with which a method for determining exposures according to the present invention is applied, there are three complementary color filters 16, 17, 18, namely cyan, magenta and yellow. Each color filter is controllably insertable independently of the other two between an illumination lamp 10 and a diffusion box 11 which comprises a square hollow tube 14 having inner mirrored walls and top and bottom diffusion plates 15. The white light from the lamp 10 passes through the color filter 16, 17 or 18 into the diffusion box 11 and is diffused thereby. The diffused light, after passing through a color original or color negative 13 placed between a negative carrier 12 and a mask member 21 so as to be held flat in an exposure position, is focussed on a color paper 26 by means of the enlarging lens 24 to create a latent image of the color original 13 under the control of a shutter 25 which is controlled by a shutter drive section 27 to open for a certain time, providing a standardized exposure time. After exposure, the color paper 26 is wound up on a take-up reel 28 and an unexposed part thereof is withdrawn from a supply reel 29 to be placed in the exposure position.

In the optical path 20 of the illumination light, a half-mirror 32 is disposed inclined by, for example in this embodiment, 45.degree. to reflect the illumination light and then to direct it to another half-mirror 34, where the light is divided into two beams, namely the light reflected toward a color TV camera 36 and the light transmitted to a sensor unit 35 comprising a lens 37, a color filter 38 and an image sensor 39. The lens 33 in front of the half-mirror 34 focuses the image of the color original 13 on the image sensor 39. From this sensor unit 35, color signals of red, green and blue can be extracted either independently by color or mixed for each picture element of the image of the color original. For example, in this embodiment, a single image sensor which is contributive to decreasing the cost of the color printer is used to detect the three color signals of the image for 100.times.100 picture elements (measuring points) thereof. It is of course permissible to replace the single image sensor by three image sensors for red, green and blue.

The respective color signals from the sensor unit 35, specifically from the image sensor 39, are transferred to an A/D converter 40 and converted thereby into digital signals which are, after being logarithmically transformed into corresponding color density signals by a logarithmic transformer 41, transferred into a microcomputer 43 through I/O port 42 thereof. This microcomputer 43 is well known per se and comprises the I/O port 42, a CPU 44, a RAM 45 and a ROM 46, so as to control the operations of components as well as to calculate exposures, as will be described later. The color density signals thus transferred are separated by color and stored in respective frame memories 45a, 45b, 45c. A keyboard 47 is provided to instruct the respective components to operate under the control the microcomputer 43.

The color TV camera 36 picks up the color negative image of the color original 13 and produces video signals. These video signals from the color TV camera 36 are transferred to a nega-posi converter 48 to be converted into positive video signals. The positive video signals thus converted are then transferred through a controller 49 to a color monitor 50 in order to display the positive image 51 on its screen. The controller 49 is adapted to produce, based on synchronizing signals, position signals which in turn are sent to a position detecting section 52.

On the other hand, an operator visually observes the positive image 51 on the screen of the monitor 50 and specifies a part of the positive image 51 which is considered to be a principal subject image from visual analysis. The operator touches the part of the principal subject image with a position-specifying means, for example, a light pen 53. This light pen 53, as is well known per se in the art, is adapted to receive light from the screen of the monitor 50 when the corresponding part of the monitor is illuminated and to photoelectrically convert the received light into a signal which in turn is sent to a position-detecting section 52. As a result, position information which represents the position of the specified part on the positive image 51 touched by the light pen 53 is given to the microcomputer 43.

As the position specifying means, it is possible to replace the light pen 53 by touch panels, joysticks, digitizers, track-balls or X-Y cursors. In addition, it is practicable to use a spot light illuminator turning on and off or a LCD panel of which a part can dim out so as to illuminate a part of the color original 13. The difference is transmitted light through a particular part of the color original 13 between the two illuminated conditions is detected by the image sensor unit 35 to detect the position of the particular part. In addition, it is possible to illuminate a spot on a principal subject image to detect the density of the spot to be compared with the base density of the color negative and then to find a picture element whose density is lower than the base density.

Reference is now had to FIG. 2 showing by way of example the image sensor 39 used in the image sensor unit 35 of FIG. 1. The image sensor 39 is comprised of a large number of photoelectric conversion elements 55, 56, 57 of the three primary colors, namely, blue, green and red which are regularly arranged in a certain order to form a matrix. Each photoelectric conversion element has a filter of corresponding color on its light-receiving surface, as is well known in the art. The photoelectric conversion elements are grouped in matrices, each of which includes three photoelectric conversion elements both per columv and per row to form a single color signal detector 58 for each picture element of the color negative image projected onto the image sensor 39. In FIG. 2, the respective single color signal detectors 58 are shown in dotted line.

When extracting color signals, color signals from photoelectric conversion elements in a single picture element are extracted additively by color, or else are extracted mixed and then added by color in the microcomputer 43, and stored in the RAM 45. In this example, it is advantageous that photoelectric conversion elements for the same colors are properly distributed in a single color signal detector 58 to eliminate the occurrence of color registration error to a considerable extent. It suffices in this image sensor 39 to arrange the color signal detectors 58, approximately 100 both per column and per row. For simplicity, it is desirable to form each color signal detector consisting of three photoelectric conversion elements arranged either in a single rank or file, or at the vertices of a triangle.

Referring now to FIG. 3, there is shown the microcomputer 43 in a functional block diagram.

Information on the coordinates of the part specified by the position specifying means 60 is transferred to a standard picture element determining section 61 in order to designate as standard a picture element included in the specified part of the principal image. In the designation of a standard picture element, if plural picture elements of the same color are included in the specified part, one at the center of the specified picture elements is designated as standard. On the other hand, if there are plural picture elements of different colors in the specified part, the designation of the standard picture element may be made in a predetermined order of priority. In the event there are plural picture elements of a color given first priority, one at the center of these picture elements is designated as standard.

The information on the position and color of the standard picture element is transferred to an image extent determining section 62 where picture elements of colors identical or similar to the color of the standard picture element are detected. The part of the image containing identical and similar colors of picture elements is designated the principal subject image; on the other hand, the remaining part of the image is designated the background image. For this purpose, picture elements are judged in succession proceeding outwardly from the standard picture element, based on their colors. At the time of judgement, it may be expedient to automatically designate as backround those parts of the image which are outside the principal subject image by a certain number of picture elements, and/or the upper part of a picture frame and/or the margins of a picture frame.

Usually, it is permissible to use a color obtained from the red, green and blue densities of the standard picture element for defining the color of the standard picture element. In this case, since the extent defining a similar color is decided with the obtained color at the center, if the obtained color is a little different from the original color of the principal subject matter photographed, the extent defining a color similar to the obtained color will deviate from the extent defining a color similar to the original color. For this reason, a principal subject image may sometimes be mistaken for a part of a background image, and the reverse applies in some cases.

In order to avoid such midjudgments, there is previously defined in the embodiment of this invention the extent defining a particular color (which is hereinafter referred to as a standard color) within which colors of a number of images to be determined as a principal subject image fall. When the color of a standard picture element falls within the standard color, the judgement as to principal subject image of color originals is made, based on the standard color.

In general, the principal subject matter to be photographed is apt to be the human figure. Therefore, it is often suitable to use as a standard color flesh color reproduced in positive images. It is also suitable to use colors which are reproduced as gray (neutral color) in positive images when the principal subject matter to be photographed is snow, buildings, stone constructions and the like.

The extents defining these standard colors can be determined by the use of color maps in a two-dimensional coordinate system the axes of which respectively represent the density differences between green and blue densities (G-B), and between red and green densities (R-G), as shown in FIG. 4. Otherwise, it is possible to determine the extent by the use of a closed steric color map in a threedimensional coordinate system whose axes respectively represent the red, green and blue color densities. When the color of a specified picture element falls within any one of these extents, it is determined to be standard.

The above-described color determination based on standard colors is carried out according to colors measured on color originals or negatives. Consequently, the colors on color originals or negatives are complementary relative to the corresponding colors to be reproduced on positives. In the following description, for understanding the present invention more easily, all colors are referred to colors reproduced on positives. The definition of flesh color and a method for judging the picture elements of flesh colors is disclosed in, for example, Japanese Patent Unexamined Publ. (Kokai) Nos. 52-156624 and 53-145620.

For determining more easily standard colors, it is suitable to determine the extent defining a particular color by the use of a rectangular color map in a two-dimensional coordinate system whose axes repectively represent the density differences between green and blue densities (G-B), and between red and green densities (R-G). In this case, the extent defining a particular color is previously determined by the coordinates of a point [(G-B).sub.0, (R-G).sub.0 ] and the distances 2.alpha., 2.beta. along the respective axes. When a color of picture element having its coordinates [(G-B).sub.1, (R-G).sub.1 ] satisfies the following absolute inequalities, the color is determined to be standard;

and

Referring again to FIG. 3, there is shown a standard color memory 63 which stores therein the data for determining the above-described extent defining standard colors. The standard picture element determining section 61 judges with reference to the data in the standard color memory 63 whether the color of the standard picture element falls within the predetermined standard color. As a result of this judgement, if in fact the color of the standard picture element falls within the predetermined standard color, the image extent determining section 62 judges, based on the result from the standard picture element determining section 61 and the data in the standard color memory 63, whether the color of each picture element falls within the predetermined standard color to determine a principal subject image by defining it with the picture elements of the colors falling within the predetermined standard color and the background image consisting of the remaining picture elements.

If the color of the standard picture element does not fall within the predetermined standard color, the image extent determining section 62 judges whether the color of each picture element is either identical or similar to the color of the standard picture element so as to determine the principal subject and background images of the color original. In this event, the extent defining similar colors is represented by color maps such as circles, rectangles, spheres or the like in two- or three-dimensional coordinate systems. Specifically, when the color coordinates of the picture element to be compared is within a color map with a center at the color coordinates of the color of the standard picture element, the color is judged to be similar to the color of the standard picture element.

Since the image extent determining section 62 determines the extents of principal subject and background images, it is possible to calculate the printing exposure for the color original based on image characteristic values extracted from the determined respective images. For example, the calculation of the red, green and blue components of printing light may be effected based either on the red, green and blue mean densities of the determined principal subject image, or on the red, green and blue mean densities of both images. In order to control the printing exposure with a high accuracy, in this invention, image characteristic values are calculated for each image of the color original and based on these the color original is classified by scene. According to the scene of the color negative, a corresponding formula is selected from different kinds of formulas for different scenes into which the color originals are classified, and to calculate the proportion of the red, green and blue components of printing light. While there are several kinds of values, considered to be such image characteristic values, red, green and blue mean color densities are used in this embodiment.

The information on positions of the images is transferred to an operation section 64 comprising mean color density calculating sections 65, 66 for principal subject images and background images, respectively, in order to calculate the mean color densities of red, green and blue for each image. Specifically, the mean color density calculating section 65 retrieves from the RAM 45 the red, green and blue densities of the respective picture elements included within the extent of the principal subject image, and then calculates the mean color density for each color, and, on the other hand, the mean color density calculating section 66 calculates the mean color density for each color of the background image in the same way as for the principal subject image. Each mean color density is obtainable either as an arithmetical mean value resulting from the total of color densities of all the picture elements divided by the number of the picture elements considered or as a median value between maximum and minimum densities.

The red, green and blue mean color densities for each image are transferred to a density difference calculating section 67 in order to calculate the gray density difference between gray densities of the two images which in turn is used as one piece of information for the scene classisification of the color original. In this section 67, a mean value or gray density is obtained from an arithmetical mean value of the mean color densities of red, green and blue for each image. From the obtained gray densities of the two images, a difference in gray density is calculated.

The mean color densities of red, green and blue of the principal subject image and the gray denstiy difference value are transferred to a calculation formula selection section 68 in which are provided various exposure calculating formulas corresponding to various kinds of scenes into which the color originals are divided. In this section 68, a scene decision is made based on the conbination of the mean color densities and the gray density difference value transferred thereto so as to select a proper formula corresponding to the kind of scene. As information for the scene decision, the gray density difference value may be replaced by either a ratio of the gray densities or a linear combination of the mean color densities of the respective images. For the exposure calculating formulas, there are known various formulas such as those disclosed in Japanese Patent Unexamined Publ. (Kokai) Nos. 52-23936 and 54-28131, and formulas using extreme (maximum and minimum) vlaues of color density, mean color density values of image areas (upper- and lower-half sections of a color original) or histogram values as the characteristic values for scene classification. It is also known to add to these exposure calculating formulas the mean color densities of the principal subject and background images of the color originals.

A characteristic value calculating section 69 calculates based on the color densities of red, green and blue of the respective picture elements stored in the RAM 45 the characteristic values and transfers them to the exposure calculating section 68. For obtaining the print exposure for each color in the exposure calculating section 68, the characteristic values are substituted in the selected formula. According to the resulting print exposure, the filter adjusting section 19 adjusts the amount of insertion of each filter 16, 17, 18 into the illumination light path 20.

FIG. 6 shows a sequence of exposure control with reference to which the operation of the color printer of FIG. 1 will be explained.

The color original or negative 13 is moved by means of a transfer mechanism (not shown) to position a picture frame to be printed between the negative carrier 12 and the mask member 21. At this time, each color filter 16, 17, 18 is retracted from the illumination light path 20 so as to allow a part of the white light from the lamp 10 to pass into the diffusion box 11 and to be diffused thereby. The diffused light, after passing through the negative 13 and the enlarging lens 24, reaches the half-mirror 32. The transmitted part of the light is focussed on the color paper 26 and, on the other hand, the reflected part of the light is directed to the other half-mirror 34 after passing through the lens 33. The half-mirror 34 divides the light incident thereto into two beams, the transmitted light being directed toward the sensor unit 35 and the reflected light toward the color TV camera 36.

The sensor unit 35 scans a color negative image focussed on the image sensor 39 therein to measure each picture element of the negative image so as to detect the red, green and blue components of light and then to provide color signals of red, green and blue for each picture element. The color signals are converted into color densities after a logarithmic transformation and then stored in the RAM 45 of the microcomputer 43 separately by color. On the other hand, the color TV camera 36 provides video signals of the negative image which are, after a negative-positive transformation, transferred to the color monitor 50 so as to display the positive image 51 of the color original 13 on its screen. A part of the positive image which is recognized through visual observation to be a part of a principal subject image is specified by use of the light pen 53. The position information of the specified part of the principal subject image is input into the microcomputer 43.

In the microcomputer 43, a standard picture element is judged as described hereinbefore and the extents or boundaries of the principal subject and its background image are determined based on the color of the standard picture element thus judged. The results of this image extent determination are written in an image extent table of the RAM 45. In FIG. 7, showing the image extent table corresponding to the negative image, each picture element 73 shown in dotted line is measured by use of a single color signal detector 58 shown in FIG. 2. When the standard picture element 74 which is shaded in FIG. 7 is determined, a rectangular judging area 75 having its center at the position of the standard picture element is established. Within the judging area, picture elements are judged as to whether their colors are identical or similar to the color of the standard picture element, so as to define a principal subject image. The reason for establishing such a judging area is to perform the determination of the principal subject image with a high accuracy, on the assumption that a principal subject image extends around the standard picture element and that parts of an image relatively far away from the standard picture element are often independent of any part of a principal subject image. Otherwise, it would be difficult to determine the extent of a principal subject image from the continuous distribution of picture elements starting from the standard picture element, without establishing a judging area.

After the establishment of the judging area 75, the standard picture element 74 is judged as to whether or not its color falls within the extent defining the standard color. When the standard picture element 74 has the standard color, each picture element in the judging area 75 is judged by comparing its color obtained based on the corresponding three color densities retrieved from the RAM 45 with the standard color whose definintion is stored in the standard color memory 63. On the other hand, if the standard picture element 74 has a color different from the standard color, each picture element in the judging area 75 is judged as to whether or not its color is identical or similar to the color of the standard picture element. In any event, the results of the judgement are written into the image extent table. As a result of this, the extent defining a principal subject image 76 which is shaded in FIG. 7 is determined and the remaining part excluded from the principal subject image 76 is determined to be a background image.

Sequentially, the image characteristic values are calculated for each image in order to make the scene classification of the negative image; and a proper exposure calculating formula is selected in accordance with the classified scene. By substituting the calculated image characteristic values into the selected formula, the calculation of printing exposure is performed by color. Based on the resulting printing exposure for each color, the filter adjusting section 19 controls each complememtary color filter 16, 17, 18 to move from its starting position toward the inside or outside of the illumination light path 20, thereby adjusting the proportion of the three color components of the printing light.

After this adjustment, by operating a print key of the keyboard 47, the shutter 25 is opened for a certain time to allow the proper exposure of the color paper 26. Upon the completion of the exposure, the color paper 26 and the color negative 13 are transported by one frame, and each color filter 16, 17, 18 is caused to return to its starting position.

In the above embodiment, only the adjustment of the color filters 16, 17, 18 is effected in order to adjust the proportion of the three color components of the printing light for obtaining prints of proper color balance and density. However, there are some kinds of color negatives for which the color filters 16, 17, 18 are not enough to make proper correction. For these kinds of negatives, it may be effective to increase or decrease the exposure time by the shutter in steps and to complementarily adjust the color filters 16, 17, 18. For improperly exposed negatives (overexposed negatives, underexposed negatives, and negatives of back-lit subject matter), the exposure calculating formulas may be used for color correction by means of the color filters 16, 17, 18 and, if so, the shutter 26 is additionally controlled to vary the exposure time for density correction.

FIG. 8 shows an example of a touch sensor panel used for the position specifying means 60.

A touch panel sensor 80 allows the user to specify a particular part of the color negative by touching it by means of a finger 81, with the result that several picture elements will be included within the touched part. In this case a sequence shown in FIG. 9 is conducted for the determination of position and color of a standard picture element.

Reference is now had to FIGS. 9 through 12 for a detailed explanation of the flow chart of FIG. 6.

As the first step, the sequence shown in FIG. 9 is executed in order to determine the position and color of the standard picture element.

Each picture element included within the part specified by the position specifying means 60 is judged as to whether or not it has flesh color. If in fact there is only one element having flesh color, that picture element is determined to be standard, and in this event, the extent of flesh color previously defined is used based on which the color judgment is made for each of the picture elements within the judging area. When there a plural picture elements of flesh color, an average one is designated as a standard picture element.

On the other hand, if there is no picture element having flesh color there among, then each picture element is judged as to whether or not it has gray color. As a result, if in fact there is one picture element of gray color, it is determined to be standard. When there are plural picture elements of gray color, an average one is designated as a standard picture element. Of course in this event the extent of gray color previously defined for the color judgment of the picture elements within the judging area is used.

Furthermore, when one picture element of gray color is present in the specified part, one of the picture elements in the specified part is designated to be standard in the same way as for flesh or gray color.

In order to determine or designate standard picture elements, it is permissible to judge each picture element based on its three color densities in place of color.

FIG. 10 shows the sequence of the color judgement for defining the extent of principal subject images.

The three color densities of the standard picture element determined or designated in the first step are retrieved, and based on the conbination thereof the color of the standard picture element is obtained. When the obtained color is either flesh color or gray color, the three color densities are retrieved for each picture element within the judging area for obtaining the color based thereon. Then the picture element is judged as to whether its color falls within the extent defining the color of the standard picture element and whether it is adjacent to a picture element which is judged to have a color falling within said extent. If the standard picture element is judged to have neither flesh color nor gray color, then each picture element is judged based on whether or not its color is identical or similar to the color of the standard picture element. From the results of this color judgment, the principal subject image of the color original is determined by the extent defined by the picture elements judged to have colors falling within the extent defining flesh color or gray color, or colors identical or similar to the colors of the standard picture element. It is also permissible for the color judgment in this sequence to compare three color densities of the standard picture element and each picture element in the judging area in place of using their colors.

FIG. 11 shows the sequence of calculating the image characteristic values and the information for the scene classification.

With reference to the results of the color judgement stored in the image area table, the color densities of the respective picture elements retrieved from the RAM 45 in order are arranged into two groups for the principal subject image and the background image, and then added by color in each group. At this time, the picture elements of each group are counted to obtain the total number thereof.

In FIG. 11, D1 designates the color density of a picture element included in the principal subject image; D2 the color density of a picture element included in the background image; i the individual picture element; and j is the color: j=1, 2 and 3 means red, green and blue, respectively. Furthermore, N1 and N2 designate the numbers of picture elements included in the principal subject and background images, respectively.

The densities are added by color and divided by the number of picture elements to provide a mean color density for each color for each image. These mean color densities are represented by D1 and D2 in FIG. 11. These means color densities are used as the image characteristic values to obtain the information for a scene classification, namely, RG (difference between the mean densities of red and green of the principal subject image), GB (difference between the mean densities of green and blue of the principal subject image), and .DELTA.D (difference between gray densities of the principal subject and background images).

FIG. 12 shows the sequence of scene classification.

According to this sequence, a first rough scene classification is made based on the tints (RG, GB) of the principal subject image, and then finely by use of .DELTA.D. Subjects which are apt to be photographed as the principal subject matter and their colors are related as follows:

flesh color (FL)--human figures

gray color (Gr)--snow, structures, stone constructions

green.about.blue (G.about.B)--scenes, plants.

According to the embodiment of the present invention, the color originals are roughly classified based on the tints of their principal subject image into four scenes of flesh color (FL), gray color (Gr), green.about.blue(G.about.B) and the others (Z), and then finely classified based on the gray color difference (.DELTA.D) into four scenes with values, .DELTA.D<-0.10, -0.10.ltoreq..DELTA.D<0.05, 0.05.ltoreq..DELTA.D<0.20 and .DELTA.D.gtoreq.0.20.

According to the classified scenes, the sixteen exposure calculating formulas F1 through F16 are selectively used. The relation between the classified scenes and the formulas is shown in Table 1.

For the above formula F1, the following general formula is used to calculate the exposure E for each color. ##EQU1## wherein the characteristic values are defined as follows: j is one of the colors red, green or blue

DFMXj is a maximum density of the whole image area

DFMIj is a minimum density of the whole image area

DCAj is a mean density of the central part of the image area

DCMXj is a maximum density of the central part of the image

DCMIj is a minimum density of the central part of the image

DLWAj is a mean desity of the lower half section of the image

DLWXj is a maximum density of the lower half section of the image

NHCNj is the number of picture elements having a density higher than 0.8.times.(DFMXj-DFMIj)+DFMIj

NLCNj is the number of picture elements having a density lower than 0.2.times.(DFMXj-DFMIj)+DFMIj

NSDj is a mean contrast of the edges of the image

NGRAY is the number or the total area of picture elements of gray color in the image.

According to the calculated exposures by color, the complementary color filters are respectively adjusted to control the proportion of the three color components of the printing light.

For the formulas F2 through F16, formulas which have the same shape as the formula F1 but have the coefficients different from each other are used according to the classified scenes of the color negatives.

It is permissible to use formulas including other kinds of image characteristic values in place of color densities.

It is also permissible to use the formulas F1 through F16 to obtain a mean exposure either for a single color or for the three primary colors. In this event, the exposure is determined on the basis of the mean exposure and the amount of color balance for each color otherwise calculated.

It may be acceptable to obtain the amount of exposure correction of the basic exposure which is determined by using, for example, the image characteristic values of the background image.

In the above description, although the scene classification if effected in accordance with the combinations of colors of the principal subject image and the gray color differences (.DELTA.D), it is not always necessary to use all the combinations listed in the table. It will be apparent to those skilled in the art that it is more advantageous to produce good prints at a high rate with a lesser number of classified scenes. Also, it may be permissible to perform the scene classification in accordance with the values of gray color difference. On the other hand, for a more accurate exposure control, it is possible to classify negatives into closely divided scenes by using as the scene classifying factor kinds of image characteristic values other than the above-described image characteristic values.

Although in the above embodiment the exposure of each color (red, green and blue) is controlled by varying the qualtiy of light in accordance with the extent of the insertion of the color filters into the optical path, it may be permissible to dispose cut-filters of cyan, magenta and yellow in the optical path when the calculated exposure for each color is given to the color paper.

Although the description has been directed to a printing apparatus to which the present invention is applied, it is to be understood that the present invention is applicable also to apparatus for examining negatives.

Claims

What is claimed is:

1. A method of detecting a principal subject image of a color original which comprises: measuring component color densities of each of a number of picture elements into which a color original is divided, said measured component color densities representing the colors of said picture elements; specifying a part of said color original by use of an image specifying means, said specified part being within a principal subject image of said color original; comparing the measured component color densities, or the color represented by said measured component color densities, of each of said picture elements with the measured component color densities, or the color represented by said measured component color densities, of one of said picture elements which is within said specified part in order to detect picture elements having a color identical or similar to the color of said one picture element; and defining the extent of said principal subject image as encompassing the detected picture elements having a color identical or similar to the color of said one picture element.

2. A method as defined in claim 1, wherein the detection of picture elements having a color identical or similar to the color of said one picture element is effected within a predetermined area surrounding said one picture element, said one picture element being designated a standard picture element.

3. A method as defined in claim 1, wherein a predetermined range of color similar to a particular color is defined, and wherein the detection of picture elements having a color identical or similar to the color of said one picture element is effected on the basis of whether or not the color of each of said picture elements falls within said range of color when at least one picture element which is within said specified part has a color falling within said range of color.

4. A method as defined in claim 3, wherein said particular color is a color in a negative original corresponding to flesh color in a positive original.

5. A method as defined in claim 4, wherein the detection of picture elements having a color identical or similar to the color of said one picture element is effected within a predetermined area surrounding said one picture element, said one picture element being designated a standard picture element.

6. A method as defined in claim 3, wherein said particular color is a color in a negative original corresponding to gray in a positive original.

7. A method of determining printing exposures for color printing by using one exposure calculating formula selected from a plurality of exposure calculating formulas in accordance with values characteristic of a principal subject image and a background image of a color original which comprises: measuring component color densities of each of a number of picture elements into which a color original is divided, said measured component color densities representing the colors of said picture elements; specifying a part of said color original by use of an image specifying means, said specified part being within a principal subject image of said color original; comparing the measured component color densities, or the color represented by said measured component color densities, of each of said picture elements with the measured component color densities, or the color represented by said measured component color densities, of one of said picture elements which is within said specified part in order to detect picture elements having a color identical or similar to the color of said one picture element; defining the extent of said principal subject image as encompassing the detected picture elements having a color identical or similar to the color of said one picture element, and defining the extent of a background image of said color original as encompassing the remainder of said picture elements into which said color original is divided; and obtaining at least one kind of value characteristic of each of said principal subject image and said background image based on the measured component color densities of the picture elements of the respective images, said obtained values being for use in selecting said one exposure calculating formula.

8. A method as defined in claim 7, wherein the detection of picture elements having a color identical or similar to the color of said one picture element is effected within a predetermined area surrounding said one picture element, said one picture element being designated a standard picture element.

9. A method as defined in claim 7, wherein a predetermined range of color similar to a particular color is defined, and wherein the detection of picture elements having a color identical or similar to the color of said one picture element is effected on the basis of whether or not the color of each of said picture elements falls within said range of color when at least one picture element which is within said specified part has a color falling within said range of color.

10. A method as defined in claim 9, wherein said particular color is a color in a negative original corresponding to flesh color in a positive original.

11. A method as defined in claim 10, wherein the detection of picture elements having a color identical or similar to the color of said one picture element is effected within a predetermined area surrounding said one picture element, said one picture element being designated a standard picture element.

12. A method as defined in claim 9, wherein said particular color is a color in a negative original corresponding to gray in a positive original.

13. A method as defined in claim 7, wherein each said at least one kind of image-characteristic value is a mean component color density.

14. A method as defined in claim 7, wherein said one exposure calculating formula is selected on the basis of the difference between gray densities of said principal subject image and said background image.

15. A method as defined in claim 7, wherein said one exposure calculating formula is selected on the basis of the difference between gray densities of said principal subject image and said background image and on the basis of mean component color densities of said principal subject image.

16. A method as defined in claim 7, further comprising: projecting printing light through said color original onto a medium on which said color original is to be reproduced, thereby effecting an exposure of said medium; disposing adjustable color filters in the path of said printing light before it passes through said color original; and controlling the position of said filters and the duration of said exposure of said medium to said printing light in accordance with said one exposure calculating formula.

Patent Citations (17)

Patent	Date	Inventor
US3677641	1972-07-01	King et al.
US4100424	1978-07-01	Akimoto et al.
US4203671	1980-05-01	Takahashi et al.
US4244653	1981-01-01	Asai et al.
US4307962	1981-12-01	Jung
US4414635	1983-11-01	Gast et al.
US4531150	1985-07-01	Amano
US4668082	1987-05-01	Terashita et al.
US4692794	1987-09-01	Suzuki
US4693596	1987-09-01	Shigaki
JPX52-23936	1977-02-01
JPX52-156624	1977-12-01
JPX53-145620	1978-12-01
JPX54-28131	1979-03-01
JPX55-29412	1980-08-01
JPX58-9136	1983-01-01
JPX59-164547	1984-09-01

Abstract

Description