Image analysis – Image enhancement or restoration – Image filter
Reexamination Certificate
1998-12-21
2003-03-25
Rogers, Scott (Department: 2724)
Image analysis
Image enhancement or restoration
Image filter
C382S275000, C358S463000
Reexamination Certificate
active
06539125
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to the art of image processing. It finds particular application in conjunction with median filters for image processing systems such as digital copiers and/or printers, and will be described with particular reference thereto. However, it is to be appreciated that the present invention is also amenable to other like applications.
In the field of image processing, a median filter is a known tool used for removing noise from an image. In operation, a typical median filter replaces each pixel intensity value with a new intensity value that is substantially equal the median intensity value of all the pixels in a surrounding neighborhood of the object pixel. Traditionally, median filters have been good at removing “salt-and-pepper” noise (i.e., erroneous values in isolated pixels) from an image. Moreover, median filters are advantageous in that they tend to preserve edges in an image. However, median filters also smooth fine textures. That is to say, for example, thin lines and other fine texture in the image are often removed by the traditional median filter. This degradation of image quality is generally unwanted in many applications where high definition image processing is desired.
The present invention contemplates a new and improved median filter and technique which overcomes the above-referenced problems and others.
SUMMARY OF THE INVENTION
In accordance with one aspect of the present invention, a method of filtering noise out of an image is provided. The image comprises pixels having varying intensity values. The method includes reading the intensity values for pixels within a neighborhood of an object pixel. A first number of intensity values are selected from the read intensity values. The selected intensity values are those that are closest to the intensity value of the object pixel. The median value for the selected intensity values is then calculated. Ultimately, the median value is assigned as a new intensity value for the object pixel.
In accordance with a more limited aspect of the present invention, prior to selecting the first number of intensity values, the read intensity values are sorted from lowest to highest.
In accordance with a more limited aspect of the present invention, the first number is such that less than all the read intensity values are selected.
In accordance with a more limited aspect of the present invention, the neighborhood is centered on the object pixel and has dimensions of N pixels by N pixels where N is a whole number.
In accordance with a more limited aspect of the present invention, the first number is less than or equal to 2N−1.
In accordance with a more limited aspect of the present invention, the first number is selectively adjustable.
In accordance with a more limited aspect of the present invention, the first number is automatically adjusted based on one or more detected characteristics of the image.
In accordance with a more limited aspect of the present invention, the one or more detected characteristics are selected from a group consisting of variance in intensity values of pixels and amounts of noise in an area of the image being filtered.
In accordance with a more limited aspect of the present invention, the steps of the method are repeated for each pixel of the image such that on each repetition a consecutive pixel of the image is designated as the object pixel.
In accordance with another aspect of the present invention, an image-processing system for filtering noise out of images while maintaining texture is provided. It includes an image-input terminal that reads an image as a plurality of pixels with varying intensity values. It also includes an image-processing unit which includes a filter. The filter: (i) samples intensity values of pixels in a neighborhood including an object pixel; (ii) selects a first number of intensity values from the sampled intensity values such that those selected are closest to an intensity value of the object pixel; (iii) calculates a median value from the selected intensity values; and (iv) assigns the median value as a new intensity value for the object pixel. Finally, an image-output terminal renders the image in accordance with the new pixel intensity values.
In accordance with a more limited aspect of the present invention, the filter is scanned across the image such that each consecutive pixel of the image is successively designated the object pixel.
In accordance with a more limited aspect of the present invention, the image-input terminal includes one or more image readers selected from a group consisting of a scanner, a stored image reader, and a computer image generator.
In accordance with a more limited aspect of the present invention, the image-output terminal includes one or more image rendering engines selected from a group consisting of an image storage device, a digital printer, a laser printer, and a video display monitor.
One advantage of the present invention is it ability to suppress “salt-and-pepper” noise in image processing systems.
Another advantage of the present invention is that while suppressing the noise image texture is preserved.
Yet another advantage of the present invention is that it may be selectively tuned prior to image processing and/or on the fly in response to certain detected image characteristics.
Still further advantages and benefits of the present invention will become apparent to those of ordinary skill in the art upon reading and understanding the following detailed description of the preferred embodiments.
REFERENCES:
patent: 4800511 (1989-01-01), Tanaka
patent: 4827533 (1989-05-01), Tanaka
patent: 5068909 (1991-11-01), Rutherford et al.
patent: 5506699 (1996-04-01), Wong
patent: 6108455 (2000-08-01), Mancuso
patent: 6208764 (2001-03-01), Archer et al.
patent: 1003332 (2000-05-01), None
“Performance Evaluation of Some Noise Reduction Methods”, By Wen-Yen Wu et al., In Graphic Models and Image Processings, vol. 54, No. 2, Mar. 1992, pp. 134-146.
Fay Sharpe Fagan Minnich & McKee LLP
Rogers Scott
LandOfFree
Modified median filter that better preserves textures does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Modified median filter that better preserves textures, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Modified median filter that better preserves textures will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3056445