Artificial Honey Bee Swarm Intelligence for the Autograding of EBN

Abstract

Swarm intelligence represents a group of computing paradigms inspired by the collective behaviour of animals and insects. While there are many swarm intelligence inspired algorithms there are those modeled after the intelligent behavior bees which have shown to be effective in doing many challenging tasks in recent years. Edible bird nest (EBN) is an expensive animal bioproduct produced by swiftlets which is deemed to have many health benefits. Currently, the common way of grading the EBN that is effective is by using trained human operators. In addition to being an unattractive job this approach is time consuming, cost ineffective with inconsistencies in the grading. Moreover, the whole time consuming training regime has to start all over again when the operator has to be replaced when they leave. In this paper we investigated the use of the artificial honey bee swarms on a set of unique features extracted by image processing to autograde EBN. An accuracy of nearly 86% was achieved on the datasets used with our proposed method.