SIGNAL-DEPENDENT NOISE CHARACTERIZATION FOR MAMMOGRAPHIC IMAGES DENOISING
Abstract
The paper deals with the noise characterization under the assumption of a heteroscedastic signal–dependent noise model in the context of medical imaging. In particular, in this kind of application, a sophisticated noise variance estimation algorithm is applied using robust estimators and nonlinear regressions. A direct relation between noise variance and pixel intensity values is obtained and used within a multiresolution denoising algorithm, performed by Wavelet Thresholding (WT). We will provide results of the noise estimation, by applying the proposed method to mammographic images.
Event details
- Event
- Exploring New Frontiers of Instrumentation and Methods for Electrical and Electronic Measurements
- Technical Committee
- TC4
- Place
- Florence, ITALY
- Time
- 22 September 2008 - 24 September 2008
- Website
- http://www.imeko2008.org/