Method and apparatus for noise reduction
First Claim
1. A method for reducing noise in a digital image represented by an array of pixel values by processing said image, said processing comprising the steps of:
- a) decomposing said image into a multiresolution representation comprising a set of detail images at multiple resolution levels and a residue,b) estimating the noise level in each detail image or recalling said noise level from a noise table comprising predetermined typical noise level values corresponding to detail images with similar noise statistics;
c) establishing a local observation neighborhood around each pixel of each detail image and estimating local image content within said neighborhood,d) pixelwise attenuating the detail images as a function of the estimated amount of image content and in accordance with said estimated noise level,e) computing said processed image by applying a reconstruction algorithm to the residue and to the attenuated detail images, said reconstruction algorithm being such if it were applied to the residual image and the detail images without attenuation, then said digital image or a close approximation thereof would be obtained.
2 Assignments
0 Petitions
Accused Products
Abstract
A noise reducing method and associated apparatus for use in a medical radiographic imaging system wherein an image represented by an array of pixels is processed and the processed image is recorded on a recording medium or visualized on a display monitor. The processing comprises the steps of
a) decomposing an original image into a sequence of detail images or into an array of coefficients representing detail strength at multiple resolution levels and a residual image,
b) pixelwise attenuating the detail images or the coefficient arrays according to the locally estimated amount of relevant signal present and in accordance with an estimated noise level,
c) reconstructing a processed image by accumulating detail obtained from the attenuated detail images or from the attenuated detail coefficients, and further adding the residual image.
-
Citations
25 Claims
-
1. A method for reducing noise in a digital image represented by an array of pixel values by processing said image, said processing comprising the steps of:
-
a) decomposing said image into a multiresolution representation comprising a set of detail images at multiple resolution levels and a residue, b) estimating the noise level in each detail image or recalling said noise level from a noise table comprising predetermined typical noise level values corresponding to detail images with similar noise statistics; c) establishing a local observation neighborhood around each pixel of each detail image and estimating local image content within said neighborhood, d) pixelwise attenuating the detail images as a function of the estimated amount of image content and in accordance with said estimated noise level, e) computing said processed image by applying a reconstruction algorithm to the residue and to the attenuated detail images, said reconstruction algorithm being such if it were applied to the residual image and the detail images without attenuation, then said digital image or a close approximation thereof would be obtained. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22)
-
-
23. An apparatus for processing an electronic representation of an image comprising
means (30) for decomposing said electronic representation into a sequence of detail images at multiple resolution levels and a residual image at a resolution level lower than the minimum of said multiple resolution levels, means (61) for storing said detail images, means (62-70) for estimating the noise level in each detail image or recalling said noise level from a noise table comprising predetermined typical noise level values corresponding to detail images with similar noise statistics; -
means (62-66) for estimating local image content within a neighborhood around each pixel in each said detail image means (71,73) for pixelwise attenuating each detail image as a function of said local image content in accordance with said estimated noise level, means (34) for computing said processed image by applying a reconstruction algorithm to the residual image and the attenuated detail images, said reconstruction algorithm being such that if it were applied to the residual image and the detail images without attenuation, then an original image or a close approximation thereof would be obtained. - View Dependent Claims (24, 25)
-
Specification