Image analysis – Image enhancement or restoration – Image filter
Reexamination Certificate
1999-12-08
2004-04-27
Wu, Jingge (Department: 2623)
Image analysis
Image enhancement or restoration
Image filter
C382S173000
Reexamination Certificate
active
06728416
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention generally relates to improving the contrast in digital images, and more specifically to an adaptive recursive filter which forms a pedestal signal from the original digital image. A tone scale function is applied to the pedestal signal and a texture signal is added to attain a processed digital image.
2. Description of the Related Art
It is well known that the dynamic range of photographic paper is less than the typical scene dynamic range. The result of this incongruity is that a good deal of scene content is rendered to black or white on the photographic print. For this reason, in an image-processing environment, a tone scale function may be used to reduce the scene dynamic range in order to map more information onto the display medium. There exist many processes for creating a tone scale function on an image dependent basis (e.g., see, U.S. Pat. No. 5,471,987 to Nakazawa et al. (hereinafter “Nakazawa”), incorporated herein by reference). Each of the conventional tone scale function processes examines certain statistical characteristics of the image under consideration in order to automatically generate the tone scale function. In addition, tone scale functions may be generated with manual interactive tools.
After the tone scale function has been generated, there exists the question of how to apply the tone scale function to the digital image. The goal of dynamic range compression is to adjust the overall dynamic range of the image, rather than to affect the contrast of any given object in the image. In essence, tone scale functions meant to reduce the image dynamic range should be applied in such a way as to minimize the effect on the scene texture. This criterion excludes the possibility of applying the tone scale function directly to the image luminance channel. Thus, it is common to apply the tone scale function to a lower frequency sub-band of the image, preserving the higher frequency sub-band(s) that are considered to be image texture (e.g., see, U.S. Pat. No. 5,012,333, to Lee et al. (hereinafter “Lee”) incorporated herein by reference).
As mentioned above, after the tone scale function has been generated, there exists the question of how to apply the tone scale function to the digital image. Application of a tone scale function meant for dynamic range compression directly to each color channel of an image results in desaturation. For this reason, it is a common practice to apply the tone scale function to a luminance (neutral) representation of the image. Direct application of the tone scale function to the image neutral channel tends to result in compression of detail in addition to compression of the overall image dynamic range, resulting in an image with a flat appearance.
Lee describes a procedure for preserving the high frequency detail of an image by blurring the image neutral channel in order to create a lowpass signal. Subtracting the lowpass signal from the image neutral channel produces a highpass signal. The processed image is generated by applying the tone scale function to the lowpass signal and adding the result to the high-pass signal. This procedure preserves a segment of the image frequency spectrum; however, artifacts are seen at large boundaries.
A. Gallagher and E. Gindele built on this work with U.S. Pat. No. 6,317,521, based on application Ser. No. 09/163,645, filed Sep. 30, 1998 (hereinafter “Gallagher”; incorporated herein by reference). More specifically, Gallagher incorporated an artifact avoidance scheme along with a single standard FIR filter to generate the texture signal. Also, in U.S. Pat. No. 5,454,044, Nakajima suggests modifying the image contrast by the formula S
proc
=S
org
+ƒ(S
us
). In Nakajima incorporated herein by reference), the low frequency image S
us
is passed through function ∂( ) which is a monotonically decreasing function. This signal is added to the original S
org
to create the processed image S
proc
.
Another example is an FIR (finite impulse response) filter based process known as homomorphic filtering (e.g., see R. Gonzalez, R. Woods,
Digital Image Processing
, Addison-Wesley Publishing Company, New York, 1992, pp. 213-218, incorporated herein by reference), which modifies the low frequencies of an image to achieve a contrast modification. In homomorphic filtering, the high frequency information is again considered to be the image texture.
In U.S. Pat. No. 5,905,817, Matama (incorporated herein by reference) describes using an IIR (infinite impulse response) filter in essentially the same framework described by Lee. The advantage to this approach is speed. In addition, by using an IIR filter, the computational requirements remain constant despite any change to the desired level of blurring.
Each of these methods of applying a tone scale function to an image channel rely on a single blurring with a linear filter. Because of this, there is an inherent size selectivity property in the tone scale function application process. Image structures that are spatially smaller than a certain size are preserved, while details larger than that size are affected by the tone scale function. In addition, the preservation of high frequencies in an image may lead to the creation of unsharp mask type artifacts (overshoot and undershoot) in the neighborhood of large edges (characteristic of occlusion boundaries or dark shadows).
In general, it was observed that larger digital filters (used to create the lowpass signal) result is a more pleasing processed image, except for the fact that the artifacts become more objectionable. Thus, the goal is to achieve greater amounts of blur without producing the overshoot artifacts at edges. Several pyramid schemes have been developed in order to achieve this goal. Because the pyramid schemes consist of multiscale representations of the same image objects, the detail size range that is preserved may be modified throughout the image.
U.S. Pat. No. 5,467,404 to Vuylsteke et al. (incorporated herein by reference) describes a method of adjusting the coefficients of a wavelet pyramid in order to modify the contrast of the image while preserving details (and producing no artifacts). In U.S. Pat. No. 5,881,181 (incorporated herein by reference), Ito describes a general multi-resolution approach intent on achieving the same goals. These methods produce satisfactory results, but require a large number of filtering operations.
Another approach to the problem of tone scale function application is to use nonlinear filtering techniques that essentially preserve edges but blur out detail. In U.S. Pat. No. 5,796,870 (incorporated herein by reference), Takeo describes a large rectangular filter, long in the direction along an edge and short in the direction across an edge. This approach reduces the artifacts at edges, but diagonal edges pose a problem. Further, Nakazawa describes using an FIR filter whose weights are determined at each pixel location, based upon the absolute value of the difference of pixel intensities between two pixels falling under the digital filter. However, this method does not account for noise in the image, and is very time consuming.
Problems to be Solved by the Invention
None of the conventional methods discussed above allows for a relatively fast filtering means that preserves details (without requiring a specific detail size range). One drawback of conventional techniques is that direct application of the tone scale function to the image neutral channel tends to result in compression of detail in addition to compression of the overall image dynamic range, resulting in an image with a flat appearance. Further, generating the processed image by applying the tone scale function to the lowpass signal and adding the result to the high-pass signal preserves a segment of the image frequency spectrum; however, produces artifacts that can be seen at large boundaries.
There is an inherent size selectivity property in the tone scale function application process. Image structures that are spatially small
Eastman Kodak Company
Woods David M.
Wu Jingge
LandOfFree
Adjusting the contrast of a digital image with an adaptive... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Adjusting the contrast of a digital image with an adaptive..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adjusting the contrast of a digital image with an adaptive... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3195723