US5594816A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Computer based digital image noise reduction method based on over-lapping planar approximation
Assignee: Eastman Kodak Co
•
Filing: 1995-08-21
•
Granted: 1997-01-14
•
Expires: 2015-08-21
Abstract
Abstract The method of the present invention reduces noise in digital photographic images by decomposing the digital image into smooth regions and edge regions and by electronically blurring the smooth regions. Smooth regions are areas lacking any sharp detail, such as blue sky. Edge regions are regions containing sharp detail, such as grass. The method operates by choosing a set of neighborhood pixel types from a matrix of pixels representing an image. For each target pixel in the image, all neighborhoods that include the target pixel are considered and a linear least squares regression is computed to fit a plane to the code values of the pixels in each neighborhood, resulting in a goodness of fit and least squares estimate for the code value at the target pixel. Finally, a noise reduced code value for the target pixel is computed as a normalized, weighted sum of the least squares estimates for the target pixel, summed over all the neighborhoods containing the target pixel. The weights are determined as a function of the goodness of fit and the neighborhood type.
Inventors
- 👤 Martin C. Kaplan
- 👤 Heemin Kwon
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 10:32
CPC Classifications
G; G06; G06T; G06T5/00; G06T5/70; G06T5/20; G06T2207/00; G06T2207/20; G06T2207/20004; G06T2207/20012; G06T2207/20172; G06T2207/20192
Claims Analysis
Total Claims: 12
Claim 1
Independent Claim
351 words
A method processing an image of a scene discernible by the human eye comprising the steps of: (a) operating a photographic image processing system which transforms a photographic image of a scene discernible by the human eye into electrical signal representative of respective picture regions of a transformed version of said photographic image, said transformed version of said photographic image, when viewed by the human eye, containing discernible noise artifacts resulting from the transformation of said photographic image into said electrical signals; and (b) processing said electrical signals to reduce noise artifacts introduced into said transformed version of said photographic image as a result of performing step (a), and thereby produce a resulting processed version of said transformed version of said photographic image which, upon being converted into an output image discernible by the human eye, will not contain such noise artifacts that are discernible to the human eye, said processing step (b) comprising the steps of: (b1) selecting a target picture region within said transformed version of said photographic image, (b2) prescribing a neighborhood of picture regions about said target picture region, (b3) for the neighborhood prescribed in step (b2), generating an electrical signal representative of a goodness of fit (x 2 ) and an estimated value of the image picture region at said target picture region in accordance with a least squares best fit plane for said prescribed neighborhood; (b4) iterating steps (b2) and (b3) for respectively different types and respectively different positions of said prescribed neighborhood; (b5) generating an output picture region for said target picture region in accordance with a summation of all estimated values of the image picture region at said target picture region with weights based upon said goodness of fit (X 2 ), and (b6) iterating steps (b1) through (b5) for all picture regions of said transformed version of said photographic image, so as to form said resulting processed version of said transformed version of said photographic image, in which said noise artifacts that have been introduced into said transformed version of said photographic image as a result of performing step (a), are reduced.
Claim 2
Independent Claim
351 words
A method processing an image of a scene discernible by the human eye comprising the steps of: (a) operating a photographic image processing system which transforms a photographic image of a scene discernible by the human eye into electrical signal representative of respective picture regions of a transformed version of said photographic image, said transformed version of said photographic image, when viewed by the human eye, containing discernible noise artifacts resulting from the transformation of said photographic image into said electrical signals; and (b) processing said electrical signals to reduce noise artifacts introduced into said transformed version of said photographic image as a result of performing step (a), and thereby produce a resulting processed version of said transformed version of said photographic image which, upon being converted into an output image discernible by the human eye, will not contain such noise artifacts that are discernible to the human eye, said processing step (b) comprising the steps of: (b1) selecting a target picture region within said transformed version of said photographic image, (b2) prescribing a neighborhood of picture regions about said target picture region, (b3) for the neighborhood prescribed in step (b2), generating an electrical signal representative of a goodness of fit (x 2 ) and an estimated value of the image picture region at said target picture region in accordance with a least squares best fit plane for said prescribed neighborhood; (b4) iterating steps (b2) and (b3) for respectively different types and respectively different positions of said prescribed neighborhood; (b5) generating an output picture region for said target picture region in accordance with a summation of all estimated values of the image picture region at said target picture region with weights based upon said goodness of fit (X 2 ), and (b6) iterating steps (b1) through (b5) for all picture regions of said transformed version of said photographic image, so as to form said resulting processed version of said transformed version of said photographic image, in which said noise artifacts that have been introduced into said transformed version of said photographic image as a result of performing step (a), are reduced.
Claim 3
Dependent Claim
52 words
A method according to claim 1, further comprising the step of: (c) applying said resulting processed version of said transformed version of said photographic image formed by step (b6) to an output image reproduction device, which generates said output image that does not contain said noise artifacts discernible to the human eye.
Claim 4
Dependent Claim
25 words
A method according to claim 1, wherein step (b4) includes iterating steps (b2) and (b3) for respectively different shapes and orientations of said prescribed neighborhood.
Claim 5
Dependent Claim
72 words
A method according to claim 1, wherein in step (b5), rather generating an output picture region for said target picture region in accordance with said summation of all estimated values of the image picture region at said target picture region with weights based upon said goodness of fit, said output image picture region for said target picture region is generated in accordance with a single estimated value associated with the largest weight.
Claim 6
Dependent Claim
25 words
A method according to claim 1, wherein said goodness of fit (X 2 ) is a measure of edge detection, texture detection, and image segmentation.
Claim 7
Dependent Claim
41 words
A method according to claim 1, wherein said target picture region has selected neighborhood sizes of: 21×21, 11×11, 5×5, 1×1, empirical weights of: W.sub.21×21 =64, W.sub.11×11 =16, W.sub.5×5 =16, W.sub.1×1 =1, and rolloff parameters of: τ.sub.21×21 =τ.sub.11×11 =τ.sub.5×5 =τ.sub.1×1 =1. 7.
Claim 8
Dependent Claim
200 words
A method according to claim 1, wherein said goodness of fit is defined as: ##EQU5## and the estimated code value of a target pixel positioned at (i,j) is: z.sub.ij.sup.kl,α =α.sub.kl +b.sub.kl.spsb.i +c.sub.klj and the computed noise reduced code value of the target pixel at position (i,j) is: ##EQU6## where the sums are over all kl,α such that pixel (i,j)εN Kl ,α and where the weighting function is: ##EQU7## wherein N kl ,α is a neighborhood of type α centered at pixel (k,l), W.sub.α is an empirically chosen weight for each neighborhood type, n.sub.α is the number of pixels in the neighborhood, ν.sub.α is the degrees of freedom for the least squares fit (ν.sub.α =n.sub.α -3), r.sub.α is an empirically chosen rolloff parameter of order unity, z ij is a code value, σ ij is the standard deviation at pixel (i,j), X kl ,α 2 is the goodness of fit for the linear least squares regression of a plane on neighborhood N kl ,α, and z ij kl ,α is the estimate of the regression at pixel location (i,j), where pixel (i,j) is some pixel in the neighborhood N kl ,α., and a,b,c are coefficients which minimize the goodness of fit.
Claim 9
Dependent Claim
56 words
A method according to claim 1, wherein step (a) comprises operating a digital photographic image-transforming device which transforms said photographic image of a scene discernible by the human eye into digital signals representative of respective digital code values of picture regions of an array of a digitally transformed version of said photographic image of a scene.
Claim 10
Independent Claim
296 words
A method processing an image of a scene discernible by the human eye comprising the steps of: (a) operating a photographic image processing system which transforms a photographic image of a scene discernible by the human eye into electrical signals representative of respective picture regions of an array of picture regions of a transformed version of said photographic image, said array of picture regions of said transformed version of said photographic image, when viewed by the human eye, containing discernible noise artifacts resulting from the transformation of said photographic image into said electrical signals; (b) processing said electrical signals to smooth and reduce noise artifacts introduced into said transformed version of said photographic image as a result of performing step (a), so as to obtain a processed version of said transformed version of said photographic image which, upon being converted into an output image discernible by the human eye, will not contain such noise artifacts that are discernible to the human eye, said processing step (b) comprising the steps of: (b1) selecting a target picture region position from said array of picture regions of said transformed version of said photographic image, (b2) prescribing a neighborhood of picture regions about said target picture region position; (b3) for the neighborhood prescribed in step (b2), generating a signal representative of the least squares best fit plane for said prescribed neighborhood; and (b4) generating an electrical signal representative of an output image picture region for said target picture region position in accordance with the least squares best fit signal generated in step (b3); and (c) operating an output image reproduction device in accordance with respective electrical signals generated in step (b4), so as to generate said output image that does not contain said noise artifacts discernible to the human eye.
Claim 11
Independent Claim
296 words
A method processing an image of a scene discernible by the human eye comprising the steps of: (a) operating a photographic image processing system which transforms a photographic image of a scene discernible by the human eye into electrical signals representative of respective picture regions of an array of picture regions of a transformed version of said photographic image, said array of picture regions of said transformed version of said photographic image, when viewed by the human eye, containing discernible noise artifacts resulting from the transformation of said photographic image into said electrical signals; (b) processing said electrical signals to smooth and reduce noise artifacts introduced into said transformed version of said photographic image as a result of performing step (a), so as to obtain a processed version of said transformed version of said photographic image which, upon being converted into an output image discernible by the human eye, will not contain such noise artifacts that are discernible to the human eye, said processing step (b) comprising the steps of: (b1) selecting a target picture region position from said array of picture regions of said transformed version of said photographic image, (b2) prescribing a neighborhood of picture regions about said target picture region position; (b3) for the neighborhood prescribed in step (b2), generating a signal representative of the least squares best fit plane for said prescribed neighborhood; and (b4) generating an electrical signal representative of an output image picture region for said target picture region position in accordance with the least squares best fit signal generated in step (b3); and (c) operating an output image reproduction device in accordance with respective electrical signals generated in step (b4), so as to generate said output image that does not contain said noise artifacts discernible to the human eye.
Claim 12
Dependent Claim
56 words
A method according to claim 9, wherein step (a) comprises operating a digital photographic image-transforming device which transforms said photographic image of a scene discernible by the human eye into digital signals representative of respective digital code values of picture regions of an array of a digitally transformed version of said photographic image of a scene.
References Cited (Backward Citations)
US4520505A
Cited
US4510618A
Cited
US4953114A
Cited
US4794531A
Cited
US4817180A
Cited
US5029227A
Cited
US4734770A
Cited
US4827533A
Cited
US4731865A
Cited
US4903316A
Cited
US5023919A
Cited
US4941192A
Cited
US5091972A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2000-01-14 | 2000-07-15 | 2001-01-14 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2004-01-14 | 2004-07-15 | 2005-01-14 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2008-01-14 | 2008-07-15 | 2009-01-14 | $7700.00 | Window Closed / Lapsed |