Multi-Label Segmentation via Residual-Driven Adaptive Regularization

TL;DR

This paper introduces a novel variational multi-label segmentation method that adaptively adjusts regularization based on residual spatial statistics, improving segmentation accuracy and reducing bias.

Contribution

It proposes a residual-driven adaptive regularization scheme within a convex optimization framework for multi-label segmentation, incorporating mutual exclusivity constraints.

Findings

Effective on synthetic and real images

Reduces regularization bias at convergence

Demonstrates improved segmentation accuracy

Abstract

We present a variational multi-label segmentation algorithm based on a robust Huber loss for both the data and the regularizer, minimized within a convex optimization framework. We introduce a novel constraint on the common areas, to bias the solution towards mutually exclusive regions. We also propose a regularization scheme that is adapted to the spatial statistics of the residual at each iteration, resulting in a varying degree of regularization being applied as the algorithm proceeds: the effect of the regularizer is strongest at initialization, and wanes as the solution increasingly fits the data. This minimizes the bias induced by the regularizer at convergence. We design an efficient convex optimization algorithm based on the alternating direction method of multipliers using the equivalent relation between the Huber function and the proximal operator of the one-norm. We empirically validate our proposed algorithm on synthetic and real images and offer an information-theoretic derivation of the cost-function that highlights the modeling choices made.