150 likes | 294 Views
Regional image segmentation using B-Spline level-set function. Author: Gábor Bernát. Scientific advisor : d r. ing. Szilágyi László. The Problem. What do you think when you see the following images:. Image Segmentation with problems like:. The segmentation itself
E N D
Regional image segmentation using B-Spline level-set function Author: Gábor Bernát Scientific advisor: dr. ing. Szilágyi László
The Problem • What do you think when you see the following images: • Image Segmentation with problems like: • The segmentation itself • Multiple independent segments • High noise levels • Low contrast • Scalability (rigidity of the contour)
The Solution Image Processing - Segmentation Signal Processing Variational Models Active Contours Osher-Sethian Level-Set Framework + topologically flexible Chan-Vese Regional Segmentation + can handle noise + low contrast Michael Unser B-Splines as Filters + scalability + easy implementation Olivier Bernard: Variational B-Spline Level Set IEEE Transactions On Image Processing, June 2009
How it works? • The Level-Set Function: • The evolution of the level-set:
Implementation • Object-oriented programming – C++ • Enclose functions and variables into a class: BSplineLevelSet • Simplify the usage for the user
Usage • From an application:
Performance • Intel 2.53 GHz CPU + 4GB RAM • Leaf Oliver Bernard (2009) – Intel 1.4 GHz CPU + 1 GB RAM • Scale 4 – 2.03 second
Key milestone check #1 • Results: • Run-time sneak peek:
To where from here? • Implementation on GPU • Improve level-set minimization time • Adapt it to various medical image problems • Tracking the left ventricle in an echocardiographic videos • Segmentation of teeth in computer tomography images
Questions ??????????????????