Image analysis – Color image processing – Pattern recognition or classification using color
Reexamination Certificate
2000-09-07
2004-11-16
Bam, Vikkram (Department: 2621)
Image analysis
Color image processing
Pattern recognition or classification using color
C382S224000
Reexamination Certificate
active
06819792
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to the art of rendering images. The invention finds application where the colorimetry or authorship device of an image is unknown.
2. Description of Related Art
Modern electronic image files contain identifying information. The identifying information is usually stored in what is called a header section of the image files. The header can identify a means with which the image was created. For example, the header can list an authoring device by manufacturer and model number as well as describe configuration settings that were used during authoring. Additionally, the header can describe a color space or gamut the authoring device operates in. The header information can be used by image processing equipment to properly interpret image data within the file and automatically ensure proper rendering. The header information is most important when an image created or adjusted for proper rendering on a first image processing system or device is rendered on a second image processing system or device. The header information is especially important when the second image rendering system is totally unaware of, and disconnected from, the first image processing system or authoring device. For example, when an image is anonymously posted and retrieved from a computer network or bulletin board, such as, for example, the INTERNET, header information is especially important. Modern image processing equipment depends on the header information in order to accurately and automatically translate or transform an image created for rendering on a first device in the first color space or gamut so that it can be properly rendered with a second device in a second color space. For example, authorship device or colorimetry header information is needed in order to properly convert an image into a device independent color space, such as, for example the CIELAB color space.
Unfortunately, not all image files contain an identifying header section. These image files, created on authoring devices that do not provide complete or correct identifying information, are referred to in this document as legacy images. When these legacy files are being processed, the image processing equipment must request operator intervention in order to properly identify and render the image. Often, the operator is presented with a list of possible sources for the data. For example, while processing a CMYK image of unknown origin, wherein C, M, Y, and K represent color separations (cyan, magenta, yellow, black), the user may be asked to pick the device the image was created for, or could have been created with, from a list including offset press standards such as Specification Web Offset Printing (SWOP) devices, JapanColor devices, Euroscale devices, and/or other devices. As another example, while processing scanned RGB files, choices may include scanners, scanned medium (e.g. photographic, lithographic), and tone scale (&ggr;) corrections.
There are problems with this image source identification technique. One problem is that often the operator does not know where the file came from. In these cases the operator is forced to guess which of the listed devices was used to create the image. Currently, after guessing at image authorship, the user must make a test print in order to verify his guess. If after examining the test print, the user determines that his guess was incorrect, another guess is made and another test print is created. Color printing can be a relatively slow process. For example, color printing is slower than black and white laser printing. Additionally, in at least some cases, the list of possible image authoring devices is quite long. Therefore, the process of guessing at authorship or colorimetry and generating test prints can be an expensive, time-consuming, aggravating, iterative process.
Recently, methods for identifying the authorship or colorimetry of legacy image files through the use of softproofing have been proposed. As used here, softproofing involves the emulation of the characteristics of possible authoring devices on a display device such as a computer CRT. A user, or operator of an image processor enters or selects a possible authoring device, for example, from a list of known authoring devices. The image processor then interprets and displays image file information in a manner that is based on the assumption that the selected authoring device is the actual authoring device. The characteristics of the selected authoring device are emulated on the display device. The image is rendered through that emulation. The user is able to evaluate a likelihood that the selected authoring device is the actual authoring device by studying the displayed image. If the selection is incorrect a new selection is made and the process is repeated.
The softproofing method eliminates the need to actually render or print the image after each authorship guess or selection. Therefore, the softproofing method can save a great deal of time. However, an inexperienced user may still be forced to select and evaluate a large number of possible authoring devices before finding a correct or acceptable authoring device or colorimetry description. For example, a list of possible authoring devices can contain hundreds of items to choose from. A complete list of possible authoring devices includes an entry for every scanner, word processor, desk top publisher, electronic prepress tool, image processor digital front end (DFE), computer display, spreadsheet editor, slide presentation editor, painting program, digital camera and image editor available.
Therefore, an automated method for reducing or eliminating the guesswork from legacy image colorimetry identification is needed. A method is needed that identifies image colorimetry so that proper image transformation and processing techniques can be applied to legacy images.
Note, in this document, the term colorimetry refers to the information that enables a relationship to be defined between the image data and an unambiguously defined color space (such as CIELAB, CIEXYZ, etc.) Identifying, for example, that a CMYK file was prepared for a SWOP press is equivalent to identifying the colorimetry of the file, since industry standard color tables exist that translate between SWOP CMYK and CIELAB.
BRIEF SUMMARY OF THE INVENTION
To those ends the present invention comprises a method for determining the colorimetry of an image file. The method comprises the step of calculating at least a colorant relationship metric based on a relationship between reciprocal colorant combinations in pixels comprising at least part of the image. Reciprocal colorant relationships occur where one colorant or set of colorants can be used to replace or substitute for some portions of another set of colorants. Additionally the method comprises the step of associating the image with one of a plurality of image classes based, at least in part, on a similarity between the first calculated metric and predetermined metrics associated with each of the plurality of image classes.
Some embodiments of the method further comprise calculating a saturation metric based on a saturation of at least part of the image and calculating a luminance metric based on a luminance of at least part of the image.
Some embodiments use at least one of the saturation and luminance metrics in the association step.
An exemplary embodiment of the invention comprises a method for determining the colorimetry of a CMYK image. The method comprises the steps of predetermining a set of colorimetry metrics describing each of a plurality of image classes, calculating a set of colorimetry metrics describing the image, and associating the image with one of the plurality of image classes based on a relationship between the predetermined set of colorimetry metrics and the calculated set of colorimetry metrics.
The step of predetermining a set of colorimetry metrics can further comprise predetermining a first metric based on an image saturation and an image luminance, and predetermining a seco
Bam Vikkram
Fay Sharpe Fagan Minnich & McKee LLP
LaRose Colin
Xerox Corporation
LandOfFree
Automatic detection of colorimetry of unknown CMYK images does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic detection of colorimetry of unknown CMYK images, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic detection of colorimetry of unknown CMYK images will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3297258