Image analysis – Image enhancement or restoration – Artifact removal or suppression
Reexamination Certificate
2000-04-18
2004-09-14
Johns, Andrew W. (Department: 2621)
Image analysis
Image enhancement or restoration
Artifact removal or suppression
Reexamination Certificate
active
06792162
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention relates to image processing and more specifically to the enhancement of digital images.
BACKGROUND OF THE INVENTION
Digital images, such as scanned photographic images, medical images, satellite images, etc. often contain undesirable noise components. For example, photographic film, such as silver-halide film, contains grains held on a substrate, such as silver-halide, as part of its physical makeup. The characteristics of these grains are a function of the type and format of the film, with faster-speed and smaller-format film typically exhibiting more graininess. It will be understood by those skilled in the art, that silver-halide film refers to photographic film that has used silver halide as the light-sensitive agent. In most existing silver-halide films, the developed silver is turned through coupling agents into a dye image, and the silver is washed out of the film. However, after the wash such films are still commonly called silver-halide films, because their action is based on silver halide.
Color film generally comprises a blue layer, a green layer, a red layer, an antihalation layer, and a transparent substrate. In developed color film, each of the blue, green, and red layers contain dyes that are used to represent blue, green, and red colors in the development process. These dyes couple to developed areas of the film that correspond to developed silver grains. For each layer, the more silver grains that develop in any given area of the image, the greater will be the resulting density of dye in that area.
Grain sizes tend to vary randomly: certain grains are small, certain grains are large, and still other grains have sizes that are in between. During the development process, the grains that were exposed to the most light are the first to develop, and other grains develop as the development process continues. Those areas in which the most grains develop for a given layer will have the greatest density of dye. Each layer of the film has its own random and unique pattern of grain makeup, so that none of the patterns of the different layers are substantially alike. A positive film or a print of an image will also have a grain pattern.
Because graininess may have a deleterious impact on the quality of a photographic image, it is conventional practice in photography to avoid graininess in photographic images as much as possible, for example by using films with slower speeds and larger formats. Unfortunately, many situations exist where it is not possible to use these kinds of film. Also, millions of films with a high degree of graininess have already been processed over the last century. Furthermore, as film ages it tends to decay physically, which can cause distortions that affect its graininess and distort color.
Digital images may be created by scanning negatives, transparencies, or printed photographic images. The quality of such a digital image depends in part on the characteristics of the film image. For example, a digital image captured from a film image will often contain traces of the grains in the film, which comprise most of the “noise” or non-image deviations from image in the digital image. Digital images also often reflect any distortions present in the film itself. The higher the resolution of the scanning, the more grain traces and distortions will be captured in the digital image.
Many scanners modify colors and densities using gamma corrections that affect the perception of graininess. For example, a digital image produced by a scanner attempting to increase shadow contrast to overcome image underexposure will produce an image with amplified shadow grain traces relative to highlight grain traces. For example, in the case of an image of a tree against the sky, the digital image may have areas of dark shadows that have more grain traces than areas of the bright sky. Similarly, the digital image may contain areas of low-detail sky with large numbers of grain traces and areas of high-detail leaves with low numbers of traces. In other words, the range of values for grain traces in these digital images tends to fluctuate widely across intensity. In addition, a scanner may blur, i.e. subdue the high frequency content, image detail or attempt to overcome this blurring by sharpening, i.e. boosting the high frequency content, image detail. In both cases the grain traces are affected in relation to frequency, along with the image. This kind of digital image can be called “un-normalized” with respect to grain traces, and it is also often visually displeasing. This is because the human eye tends to focus on the areas in an image that have the highest amount of grain traces and to visually associate those same amounts to areas that actually have fewer grain traces.
Although one can attempt to improve the quality of a scanned image, such improvements can often only be obtained through labor-intensive alterations, which require a person of relatively high skill to subjectively determine what characteristics should be altered to improve the quality of the final image. As a result, because each image may be individually operated upon, the improvement process is time consuming and improvement of a large number of images may be impractical. Accordingly, another method of enhancing digital images is needed that can enhance a digital image automatically.
SUMMARY OF THE INVENTION
Briefly summarized, one aspect of the invention is a method of automatically enhancing a digital image. The magnitudes of grain traces are measured in at least two segments in the digital negative and the digital negative is enhanced using the magnitudes of the grain traces measured in the at least two segments.
The invention has several important technical advantages. Various embodiments of the invention may have none, one, some, or all of these advantages as well as additional advantages. The invention can be used to automatically enhance a digital image with respect to grain traces (noise). Such enhancement may be desirable for use with scanned photographic images. The invention allows automatic calculation of correction arrays for the digital image to correct uneven grain trace distribution with respect to intensity and frequency. Such arrays may be useful to normalize the grain traces in relation to pixel intensities and/or frequency. Once grain traces have been normalized, the invention allows reduction of the grain to a level chosen automatically or by a user of the invention, as desired. In addition the normalized grain trace data may be used to balance the digital image's color channels.
REFERENCES:
patent: 4677465 (1987-06-01), Alkofer
patent: 4937720 (1990-06-01), Kirchberg
patent: 5010401 (1991-04-01), Murakami et al.
patent: 5047968 (1991-09-01), Carrington et al.
patent: 5291286 (1994-03-01), Murakami et al.
patent: 5561611 (1996-10-01), Avinash
patent: 5565931 (1996-10-01), Girod
patent: 5729631 (1998-03-01), Wober et al.
patent: 6055340 (2000-04-01), Nagao
patent: 6373992 (2002-04-01), Nagao
patent: 6415053 (2002-07-01), Norimatsu
patent: 6667815 (2003-12-01), Nagao
patent: 0 569 142 (1993-11-01), None
patent: 0 816 833 (1998-01-01), None
patent: WO 90/01240 (1990-02-01), None
patent: WO 91/09493 (1991-06-01), None
patent: WO 98/31142 (1998-07-01), None
M. Doyle, et al., “Adaptive Fourier Threshold Filtering: A Method to Reduce Noise and Incoherent Artifacts in High Resolution Cardiac Images”,Magnetic Resonance in Medicine, vol. 31, No. 5, pp. 546-550.
B. Aiazzi, et al., “A Robust Method for Parameter Estimation of Signal-Dependent Noise Models in Digital Images”, IEEE, pp. 601-604, Jul. 2, 1977.
Aggelos K. Katsaggelos, et al., “A Regularized Iterative Image Restoration Algorithm”,IEEE Transactions On Signal Processing, vol. 39, No. 4, pp. 914-929, Apr. 1991.
XP-000780652: Til Aach, et al., “Anisotropic Spectral Magnitude Estimation Filters for Noise Reduction and Image Enhancement”,IEEE Proc. Int. Cont. on Image Processing, pp. 335-338, Sep. 16, 1996.
Rangaraj M. Rangayyan, et al., “Adaptive-neighborhood filtering of image
Baker & Botts
Eastman Kodak Company
Johns Andrew W.
Novais David A.
LandOfFree
Method and apparatus to automatically enhance the quality of... 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 to automatically enhance the quality of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus to automatically enhance the quality of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3227152