Image analysis – Image sensing
Reexamination Certificate
1999-02-18
2002-10-01
Grant, II, Jerome (Department: 2722)
Image analysis
Image sensing
C358S504000, C382S155000
Reexamination Certificate
active
06459825
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATIONS
Not applicable.
BACKGROUND
1. FIELD OF INVENTION
The invention relates to apparatus and methods of automated photo image scanning for photo reprographics from various slide transparency, negative film, and reflective print media. More particularly, the invention is directly applicable to demanding photo-optical applications where scanner equipment is required to image capture at the highest quality image resolution and fidelity possible.
2. DESCRIPTION OF RELATED ART
Analog film with its infinite color representation and optical emulsion continuous tone fidelity has been traditionally an extremely good source for high quality image reproduction. This work, historically, was performed by analog photo optical techniques. For instance, one extreme example is the case of film placed in a large optical projection camera, which is then constructed into a large darkroom, and optically exposing to large format photo paper. This kind of application for film has for years been used to create many of today's photo KIOSKS or large format photo advertising banners. Other high quality photography and film applications have included school and portrait photography, pre press, motion pictures, and thousands of other professional uses of film as tooling for image reproduction.
It has always been an industry difficulty, even before the advent of digital imaging, to obtain accurate photo reproduction color and optical quality. This gave rise to the use of video cameras and scanners as colorimeters for determining the corrected values to be used in the analog reproduction process. U.S. Pat. No. 3,893,166 to Pugsley (1975). Color correcting image reproduction method and apparatus.
With the advent and proliferation in recent years of digital imaging output devices, the demand for the higher quality and speed of digital scanning has grown. In general, the tools have not been up to the quality standards for many of the tasks. The highest state of the art has been for a skilled scanner operator to use personal judgement to scan film positive transparencies (slide film). The operator interacts on a computer workstation, with a color monitor, to select scanner parameters and create digital image data files. U.S. Pat. No. 5,155,588 to Levien (1992) discloses a color correction and apparatus for photographic reproduction.
The limitations of this approach are due to many factors. First, the entire approach depends on human judgement, utilizing an inaccurate color monitor to subjectively balance color and set resolution and optical control parameters. The actual quality results will not be totally visible until after printing the media. Additionally, due to the combination of the density to sensitivities of the film with the intensity variations of a scanner, the highest quality processes are limited to only a bump function and only applicable to positive transparency film.
Thereafter, inventors created calibration techniques for color calibrating scanners to attempt to compensate for exposure control with test reflective photography. U.S. Pat. No. 5,721,811 to Eckhardt (1998) discloses a pre press system color control technique which can only scan slide transparencies utilizing an instant color photograph as a photo color reference for lighting conditions. This inventor states that photographic color film transparencies are exclusively used because using negatives with color would add complexity, loss of sharpness, and color distortion to the process. Photo prints are generally inadequate because the image print resolution quality is much lower than film. Negative film is not often used because the scanner and computer equipment is not up to the task. More than 90% of all film sold and used is negative film, yet none of these materials are currently suitable for the high quality digital image reproduction process because of scanning limitations. Negative film densities are reversed, offset in yellow, magenta, and cyan dye densities, and very compressed between the representatives of information content between black and white. Typical spectral density distribution compression is between 35 and 50%, (depending on the type of negative film), making digitizing negative film very sensitive to computer error.
Computer scanners are typically manufactured with log curves built into the equipment in order to approximate (only) film response and to display attractive results to video gamma devices including computer monitors. However, all photo film dye densities are characterized not by log curves, but more generally “S” shaped curves that are unique to each material, with shadow and highlight detail fall off Thus, spectral response shadow and highlight details are damaged or lost with log curve scanning.
The digital fidelity scanning color depth process, to the computer from the scanner, is typically controlled by a 24-bit to 36-bit color lookup table (analog to digital converter), and a 24-bit color depth (16.7 million available colors). The color spectral compression of negative film densities has the effect of reducing the scanner intensity data space to be reduced by 35%-50% of the digital domain of the scanner. Therefore, the 24-bit data space is reduced to representative values of between approximately 16 to 12-bit. Thus, the available photo image color space is reduced from the infinite number available on the analog negative film, and the 16.7 million potential colors of the scanner, to the lowest common film to scanner intensity response of the film density. That is between 65,535 to 4096 individual digital color representation for 16 to 12-bit respectively. This often causes visible posterization or duplications of colors in digital image scans of negative film.
Further complicating the photo-media scanner technology state of the art is the fact that not only is every film stock media different in the dye or ink stock that it is formulated with, but also each photographer's individual exposure of the film can have slight differences in light levels. With highly sensitive films (especially negatives) the color correction curves for each different exposure varies significantly. Therefore, standardized color corrections are not enough to avoid the problems of serious color shifts and aberrations in the results. U.S. Pat. No. 5,475,493 to Yamana (1995) discloses a gray balance correcting method to address this problem. The limitation of this approach is that it is based solely on three points, white/Dmin, black/Dmax, and gray densities. Due to the digital domain data sensitivity issues previously mentioned, it is necessary to have all the data values possible, both to avoid color distortion and to digitally capture to support the full fidelity of the film's spectral data.
Two types of image processing generalized analysis methods have been proposed. The first, U.S. Pat. No. 5,200,816 to Rose (1993), provides for color processing with learned neural networks based on utilizing KODAK Q60 Targets as a scanner reference. This proposal, based on its dependency on the ISO IT8 sub-committee target standards, as a result lacks all reference to primary colors of red, green, blue, magenta, yellow, and cyan between black and their fully saturated color representation. Additionally, neither KODAK nor other sources provide this target on any photo media except on limited varieties of positive (SLIDE) transparencies and print paper. Thus, the referenced learning color method lacks access and methodology for the color fidelity information necessary for the most demanding applications and is not applicable to negative film altogether.
The second, U.S. Pat. No. 5,748,773 to Tashiro (1998), is a laser copy machine pre-scan operation that detects the material type by looking for a predetermined set of color feature points from the formed histogram of the pre-scan. By comparing the sampled feature points to a second set of feature points, it then converts the data to a third color feature points set. Although pre-sampling the media and performing a histogram analysis give
LandOfFree
Method and apparatus for a self learning automatic control... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for a self learning automatic control..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for a self learning automatic control... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2954103