Image analysis – Color image processing
Reexamination Certificate
2000-03-03
2003-12-16
Johns, Andrew W. (Department: 2621)
Image analysis
Color image processing
C382S167000, C382S275000, C358S518000
Reexamination Certificate
active
06665434
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image processing device, an image processing method, and a recording medium, and particularly to an image recording method for correcting, in image data representing an original image formed by recording an object by photography, at least the color imbalance of the original image caused by the light source when the object was recorded by photography, an image processing device which can make appropriate use of the image processing method, and a recording medium for recording a program which enables a computer to function as the image processing device.
2. Description of the Related Art
The color balance of an image visualized on a photographic film by performing developing processing and the like on the photographic film on which an object has been recorded by photography with a camera or the like (i.e. of an image recorded by photography on photographic film), is affected by the characteristics of the photographic film itself and by the processing conditions during the image processing (these will collectively be referred to below as “film characteristics”), and the color balance may become biased compared to the color balance of the object when photographed. For this reason, when an image recorded by photography on a photographic film is recorded on a recording material such as photographic paper or is displayed on a display means such as a display unit, it is sometimes necessary to correct the color balance of the image recorded on the recording material or displayed on the display means (referred to below as the “output image”), so that the color balance of the output image matches the color balance of the object when photographed (i.e. so that the gray portions of the object at the time of photography are reproduced as gray on the output image).
A known example of a method for correcting the color balance of an output image is one in which, assuming that the color of the pixel corresponding to the highlight (for example, the pixel having the maximum density in a negative image, and the pixel having the minimum density in a positive image) of an image recorded by photography on a photographic film (an original image), and the color of the pixel corresponding to the shadow (for example, the pixel having the minimum density in a negative image, and the pixel having the maximum density in a positive image) of the original image are taken as white and black respectively, the gray balance, which represents the color balance on an original image of the gray portions of the photographed object at the time of photography, is estimated and the color balance of the output image is then corrected using the estimated gray balance as a reference. In this method, the pixel corresponding to the highlight and the pixel corresponding to the shadow are each extracted from the original image, and an axis joining in a straight line the points corresponding to both pixels using, for example, RGB density coordinates is determined as the gray axis representing the gray balance.
However, in the above correction method, when the color of the pixel corresponding to the highlight in an original image is not white, such as when the highlight is the forehead of the face of, for example, a Japanese person in an image in which an electronic flash has been used to photograph the person, it is not possible to determine the gray axis representing the proper gray balance (likewise with the shadow), and the color balance of the output image is biased towards the color which complements the color of the pixel corresponding to the highlight. This is known as “highlight failure”. The problem has thus been that while the appearance frequency of original images in which this type of highlight failure occurs easily is relatively high, so that likelihood of obtaining an output image with the proper color balance is low.
Another method is known in which, on the basis of Evan's theory, an image having a uniform LATD (Light Accumulation Transmission Density) for each component color (e.g. RGB), is regarded as being an image in which the color balance matches the color balance of the photographed object at the time of photography (an image in which the gray balance can be determined), the LATD (Light Accumulation Transmission Density) for each component color (e.g. RGB) of the original image is then measured, and the color balance of the output image is corrected so that the LATD for each component color of the output image is uniform.
However, in the above correction method, when regions which are non-gray and in which the hue is substantially uniform (for example, regions corresponding to a green lawn or a blue sky or sea) occupy relatively large portion of the original image, then correcting the LATD so that the LATD is uniform for each component color of the output image results in the color balance of the output image becoming biased towards the color which complements the color of the above regions This is known as color failure. Because the appearance frequency of original images in which this type of color failure occurs easily is relatively high, the problem has been that likelihood of obtaining an output image with the proper color balance is low, as is the case with the method described above.
Japanese Patent Application Laid-Open (JP-A) No. 9-83825 discloses a method in which image data DR′, DG′, and DB′ relating to low saturation pixels is acquired from image data DR, DG, and DB representing an original image, the shadow points DRs, DGs, and DBs and the highlight points DRh, DGh, and DBh are determined for the respective image data DR′, DG′, and DB′, the image data set (DR″, DG″) is obtained by, for each same values of one of a pair of the image data (DR′, DG′) corresponding to each other for each pixel, averaging the values of the other one of the pair (DR′, DG′) corresponding to the same values, the relationship between the density of the two colors is obtained from the set of the image data (DR″, DG″) and the shadow points (DRs, DGs) and highlight points (DRh, DGh), and, on the basis of the relationship thus obtained, at least one of the image data DR and image data DG undergoes linear conversion across the entire area thereof so as to be made equal with the other one of DR and DG.
In the technology described in the above publication, an attempt is made to alleviate the effects brought about when the image is one in which color failure easily occurs by removing from the data to be calculated the data of high saturation pixels from among the image data representing the original image. In addition to the high saturation pixels, pixels which are adjacent to the high saturation pixels and whose hue is different to the high saturation pixels by a predetermined amount or less are also removed from the data to be calculated. However, even if the above processing is actually carried out on an original image in which color failure is likely to occur, many pixels out of those non-gray and substantially uniformly-hued pixels which occupy a relatively large surface portion of the original image are not removed from the data to be calculated and are left remaining. Accordingly, when the original image is one in which color failure is likely to occur, there has been a problem that the color balance of the output image has not been corrected with sufficient accuracy.
Moreover, it is well-known that the color balance of images obtained by photographing and recording on photographic film scenes which are lit by different types of light sources such as fluorescent lights or tungsten lights (alternative light source scene images) is greatly affected by the fact that the spectral distribution of illuminating light from the alternative light sources is hugely different from the spectral distribution of typical light sources such as daylight and the like, and the color balance becomes biased towards a particular color relative to the original color
Alavi Amir
Birch & Stewart Kolasch & Birch, LLP
Fuji Photo Film Co. , Ltd.
Johns Andrew W.
LandOfFree
Device, method, and recordium for correcting color imbalance... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Device, method, and recordium for correcting color imbalance..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Device, method, and recordium for correcting color imbalance... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3131606