Pottslab - Multilabel segmentation of color images

Multilabel image segmentation for vector-valued images based on the Potts model

Multilabel image segmentation for vector-valued images based on the Potts model (also called piecewise constant Mumford-Shah model)
Features:
- Multilabel image segmentation (2D domain) or step detection for signals (1D domain)
- No label discretization required (labels are chosen automatically)
- Supports multichannel images (e.g. RGB, multispectral or feature images) and has linear complexity in number of channels
- Supports indirect measurements by a linear operator (e.g. blurred data, computed tomography data)
- Supports multithreading