Jackfruit Phytophthora Palmivora (Butler) Disease Recognizer Using Naïve Bayes Classifier
Authors: Jonah Flor V. Oraño, Jomari Joseph A. Barrera, and Elmer A. Maravillas,
Abstract
This paper presents a technique to detect the presence of P. palmivora disease in jackfruit trunk using Naïve Bayes classifier. In this study, 200 sample images of jackfruit trunk were used, which were divided into two sets: for training and for testing. Each set contains 50 images for healthy and 50 images for disease infected. The input images were subjected to image pre-processing such as cropping, scaling, and brightness and contrast adjustment. Then, the images were segmented into two regions using color masking. Texture features such as angular second moment (uniformity) and sum of squares (variance) were also extracted from the images. Next, Naïve Bayes classifier was used to classify whether the jackfruit is infected with the disease or not. Finally, the performance of the classifier was evaluated by computing the overall accuracy of the system. Based on the result, the classifier achieved 94% accuracy in detecting the disease incidence. Moreover, this rate can be further improved by adding texture features and by applying other classification algorithms.