Evolutionary optimization applied for fine-tuning parameter estimation in optical flow-based environments

Abstract

Optical flow methods are accurate algorithms for estimating the displacement and velocity fields of objects in a wide variety of applications, being their performance dependent on the configuration of a set of parameters. Since there is a lack of research that aims to automatically tune such parameters, in this work we have proposed an evolutionary-based framework for such task, thus introducing three techniques for such purpose: Particle Swarm Optimization, Harmony Search and Social-Spider Optimization. The proposed framework has been compared against with the well-known Large Displacement Optical Flow approach, obtaining the best results in three out eight image sequences provided by a public dataset. Additionally, the proposed framework can be used with any other optimization technique.