MRBSiZer is a method that aims to capture scale-dependent features in a random signal. The method is based on scale space smooothing. However, while the usual scale space analysis approach is to suppress detail by increasing smoothing progressively, the proposed method instead considers differences of smooths at neighboring scales. A random signal can then be represented as a sum of such differeces, a kind of a multiresolution analysis, each difference representing details relevant at a particular scale or resolution. Bayesian analysis is used to infer which details are credible and which are just artifacts of random variation.

For more information see article
Scale space multiresolution analysis of random signals
Holmström L, Pasanen L, Furrer R, Sain S
Computational Statistics and Data Analysis 55 (2011) 2840-2855.


MATLAB-Routines are written using MATLAB 2008b. Some of the routines use MATLAB Statistics and Image processing toolboxes.

Since we process images, the computer used should have a rather large RAM-memory. Due to the large size of the variables some effort has been put into to making the code efficient. Still, the author makes no claims as to the optimality of the code. Even some bugs might still be found!


Download all MRBSiZer-functions and subroutines in a one zip-file.

11.02.2013: A bug fixed from MRBSiZer_Sphere.m and MRBSiZer.m.
MRBSiZer.m is also now more efficient.

Examples of how to use the MATLAB-commands can be found in the Interface section. Subroutines can be viewed separately and linked to the interfaces in the Subroutines section. The test images can be downloaded from Test Images section.


This software is free only for academic and personal use. It must not be modified and distributed without prior permission of the author. The author is not responsible for implications from the use of this software.