A First Derivative Potts Model for Segmentation and Denoising Using ILP
This addresses a combined image processing problem for computer vision applications, but appears incremental as it builds on existing Potts and ILP methods.
The paper tackles unsupervised image segmentation and denoising simultaneously by proposing a novel integer linear programming (ILP) formulation of the first derivative Potts model with an ℓ₁ data term, using binary variables to handle the ℓ₀ norm, and solves it with a standard MIP solver, showing results compared to the multicut problem.
Unsupervised image segmentation and denoising are two fundamental tasks in image processing. Usually, graph based models such as multicut are used for segmentation and variational models are employed for denoising. Our approach addresses both problems at the same time. We propose a novel ILP formulation of the first derivative Potts model with the $\ell_1$ data term, where binary variables are introduced to deal with the $\ell_0$ norm of the regularization term. The ILP is then solved by a standard off-the-shelf MIP solver. Numerical experiments are compared with the multicut problem.