The Honeybee Search Algorithm: A Cooperative Coevolutionary Framework for 3D Reconstruction

Abstract

This chapter describes a novel analogy that mimics the way in which honeybee colonies work, with the goal of approaching the problem of sparse and quasidense reconstruction. Here, a new adaptive behavior strategy is presented based on the “divide and conquer” strategy employed by the honeybee colony for approaching search problems. This work investigates the communication system of honeybees with the purpose of obtaining an intelligent approach for three-dimensional reconstruction. A novel framework is proposed in which the 3D points communicate among themselves to achieve an improved sparse reconstruction which could be used reliably in further visual computing tasks. The general ideas that explain honeybee behavior are translated into a computational algorithm following the paradigm of evolutionary computing. Experiments demonstrate the importance of the proposed communication system for reducing dramatically the number of outliers in a real working system.

“There is nothing more difficult for a truly creative painter than to paint a rose, because before he can do so he has first to forget all the roses that were ever painted.”

– Henri Matisse

“The bees learn where they live by landmarks. If they’re moved within their home range, they get confused.”

– Gene Robinson