This toolbox solves the computational tomography problem where the object is partially discrete in nature, that means - there is a variation part, called background, where properties varies between 0 and tau < 1, and an anomaly of constant value 1. This kind of objects occur in many applications, from image processing to medical imaging. We use parametric level-set method to represent the anomaly and the smooth regularization (i.e. tikhonov) for background. We compare the proposed method with Total-variation method, DART and P-DART.
This toolbox requires Matlab version R2016a or later. In particular we require following packages to be pre-install to run this toolbox.
Please run startup.m file before running the scripts in example folder.
The theory and experiment details are available in the paper "A parametric level-set method for partially discrete tomography". The paper is available on ArXiv : arXiv:1704.00568
Feel free to contact me with any questions, suggestions, etc.
Ajinkya Kadu - [email protected]