Analysing the quality of phaseolus vulgaris family of legumes using artificial neural network and “bag of features” techniques

Abstract

The Philippine Council for Agriculture, Forestry and National Resources Research and Development-Department of Science and Technology (PCAARRD-DOST) have recognized the importance of cultivating legumes as priority crop among others in the vegetable industry under the National Vegetable Research & Development Program. They have further emphasized the need for innovating the methods to improve the processes in terms of producing better quality of products. The study developed a prototype compiled application based on the trained and validated dataset using ANN (Artificial Neural Network) machine. The BoF (Bag of Features) technique was utilized for image features extraction in the SVM (Support Vector Machine) environment for quality classification of Phaseolus Vulgaris family of legumes. These are commonly cultivated in the Philippines. The combined methods yielded an accuracy of 90.2%.