Automatically selecting inference algorithms for discrete energy minimisation

TL;DR

This paper introduces a method to automatically select the most effective inference algorithm for discrete energy minimization problems in computer vision, significantly improving solution quality by choosing the best algorithm for each model.

Contribution

We propose a novel technique that automatically predicts the optimal inference algorithm for a given graphical model, advancing beyond prior surveys and manual selection methods.

Findings

Achieves 96% of variables matching the best algorithm's labelling.

Validated on an extended OpenGM2 benchmark with diverse vision problems.

Outperforms manual selection in inference accuracy.

Abstract

Minimisation of discrete energies defined over factors is an important problem in computer vision, and a vast number of MAP inference algorithms have been proposed. Different inference algorithms perform better on factor graph models (GMs) from different underlying problem classes, and in general it is difficult to know which algorithm will yield the lowest energy for a given GM. To mitigate this difficulty, survey papers advise the practitioner on what algorithms perform well on what classes of models. We take the next step forward, and present a technique to automatically select the best inference algorithm for an input GM. We validate our method experimentally on an extended version of the OpenGM2 benchmark, containing a diverse set of vision problems. On average, our method selects an inference algorithm yielding labellings with 96% of variables the same as the best available algorithm.