System for manipulating noise in digital images
First Claim
1. A method of analyzing noise content in a digital image, the method comprising:
- identifying regions of approximately constant pixel value within the digital image;
selecting a predetermined number of said identified regions based on pixel value variances within each said identified region;
analyzing said selected regions to generate a mathematical model of noise present in the digital image, wherein said analyzing said selected regions to generate a mathematical model of the noise present in the digital image comprises;
processing each selected region to remove a mean pixel value from each pixel in the selected region and obtain an approximation of the noise in each selected region,analyzing the power spectral density (PSD) of the noise in said selected regions,generating a parametric model of the noise present in the digital image based on said PSD of the noise in said selected regions, wherein said generating a parametric model of the noise present in the digital image based on said PSD for each selected region comprises;
taking a discrete Fourier transform of each noise sample, andsquaring each resulting frequency component to produce a periodogram,averaging together the periodograms; and
selecting a best-fit linear combination of a small set of predefined PSDs.
4 Assignments
0 Petitions
Accused Products
Abstract
An apparatus for analyzing the broadband noise content of a digital image comprising the following: a means of automatically identifying regions of originally constant color in the image by analysis of the variance of pixel values of regions of the image; a means of automatically detecting and discarding regions deemed to be unrepresentative of the true noise content of an image, including under- and over-exposed regions; a means of allowing the user to manually select some or all required constant color regions if desired; and, a means of analyzing such constant color regions to generate a parametric or non-parametric model of the noise in the image, including frequency characteristic within and between channels and other characteristics such as phase which might describe structured noise.
-
Citations
40 Claims
-
1. A method of analyzing noise content in a digital image, the method comprising:
-
identifying regions of approximately constant pixel value within the digital image; selecting a predetermined number of said identified regions based on pixel value variances within each said identified region; analyzing said selected regions to generate a mathematical model of noise present in the digital image, wherein said analyzing said selected regions to generate a mathematical model of the noise present in the digital image comprises; processing each selected region to remove a mean pixel value from each pixel in the selected region and obtain an approximation of the noise in each selected region, analyzing the power spectral density (PSD) of the noise in said selected regions, generating a parametric model of the noise present in the digital image based on said PSD of the noise in said selected regions, wherein said generating a parametric model of the noise present in the digital image based on said PSD for each selected region comprises; taking a discrete Fourier transform of each noise sample, and squaring each resulting frequency component to produce a periodogram, averaging together the periodograms; and selecting a best-fit linear combination of a small set of predefined PSDs. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A computer readable medium containing computer executable instructions for analyzing noise content in a digital image, which, when operating in a processor cause the processor to perform the functions of:
-
identifying regions of approximately constant pixel value within the digital image; selecting a predetermined number of said identified regions based on pixel value variances within each said identified region; analyzing said selected regions to generate a mathematical model of noise present in the digital image, wherein said analyzing said selected regions to generate a mathematical model of the noise present in the digital image comprises; processing each selected region to remove a mean pixel value from each pixel in the selected region and obtain an approximation of the noise in each selected region, analyzing the power spectral density (PSD) of the noise in said selected regions, generating a parametric model of the noise present in the digital image based on said PSD of the noise in said selected regions, wherein said generating a parametric model of the noise present in the digital image based on said PSD for each selected region comprises; taking a discrete Fourier transform of each noise sample, and squaring each resulting frequency component to produce a periodogram, averaging together the periodograms; and selecting a best-fit linear combination of a small set of predefined PSDs. - View Dependent Claims (17, 18, 19, 20, 21, 22)
-
-
23. A software program product for use with a processor, the software program product comprising processor readable program code for analyzing noise content in a digital image, wherein when said processor readable program code is executed in a processor, said processor readable program code causes said processor to perform the functions of:
-
identifying regions of approximately constant pixel value within the digital image; selecting a predetermined number of said identified regions based on pixel value variances within each said identified region; analyzing said selected regions to generate a mathematical model of noise present in the digital image, wherein said analyzing said selected regions to generate a mathematical model of the noise present in the digital image comprises; processing each selected region to remove a mean pixel value from each pixel in the selected region and obtain an approximation of the noise in each selected region, analyzing the power spectral density (PSD) of the noise in said selected regions, generating a parametric model of the noise present in the digital image based on said PSD of the noise in said selected regions, wherein said generating a parametric model of the noise present in the digital image based on said PSD for each selected region comprises; taking a discrete Fourier transform of each noise sample, and squaring each resulting frequency component to produce a periodogram, averaging together the periodograms; and selecting a best-fit linear combination of a small set of predefined PSI)s. - View Dependent Claims (24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36)
-
-
37. A method of analyzing noise content in a digital image, the method comprising:
-
identifying regions of approximately constant pixel value within the digital image, wherein said identifying comprises; defining a plurality of regions in the digital image; discarding any of said defined regions in which pixel values are saturated or clipped to generate a set of remaining regions; analyzing said remaining regions to determine variation among pixel value within each remaining region; and identifying said remaining regions having lower variation among pixel value; selecting a predetermined number of said identified regions based on pixel value variances within each said identified region; analyzing said selected regions to generate a mathematical model of noise present in the digital image, wherein said analyzing said selected regions to generate a mathematical model of the noise present in the digital image comprises; processing each selected region to remove a mean pixel value from each pixel in the selected region and obtain an approximation of the noise in each selected region, analyzing the power spectral density (PSD) of the noise in said selected regions, generating a parametric model of the noise present in the digital image based on said PSD of the noise in said selected regions, wherein said generating a parametric model of the noise present in the digital image based on said PSD for each selected region comprises; taking a discrete Fourier transform of each noise sample, and squaring each resulting frequency component to produce a periodogram, averaging together the periodograms; and selecting a best-fit linear combination of a small set of predefined PSDs. - View Dependent Claims (38, 39, 40)
-
Specification