A Novel Flower Pollination Algorithm for Auto-Grading of Edible Birds Nest

Abstract

Edible Bird Nest (EBN) produced by certain species of swiftlets has been known of its source of protein and vitamins that benefit the human body. This results in high demand from humanity due to the advantages of consuming the EBN. However, manual process of grading and classifying the EBN for different price range may cause drawbacks towards the production of EBN. The grading of EBN is done by observing the colour, shape, size and impurities present in the nest. Although manual process is done by trained personnel, the results obtained are often inconsistent and inaccurate due to human fatigue. Hence, this process is tedious and time consuming which may cause delay in the production of EBN. To overcome this issue, a novel Drunken Flower Pollination Algorithm (DFPA) is developed to perform auto grading on the EBN. This DFPA is also compared with the existing FPA and four other popular heuristics where the DFPA achieved better grading accuracy with an average accuracy of nearly 88%.