Independent Gaussian Gray Level Geometric Neural Network Classifier For Plant Leaf Disease Prediction

Abstract

Swift and precise plant leaf disease prediction is dangerous to surging agricultural fertility within a feasible manner.
Conventionally, human experts have been reckoned upon to identify variation in plants leaf because of diseases,
pests, or paramount climate. But, the excessive, laborious during certain scenario found as unrealistic. To address
such issues, application of image processing methods to detect plant leaf disease is a more familiar topic. Therefore,
Independent Gaussian Gray Level Geometric Neural Network Classifier (IGGL-GNNC) for plant leaf disease prediction
is proposed. This IGGL-GNNC method is split into three sections. They are image enhancement and pre-processing
using Independent Component Gaussian Median Preprocessing model. Second, feature extraction using preprocessed plant leaf images by means of Gamma Corrected Gray Level Run Length Feature Extraction model. Finally,
Sine Cosine Position Update Predictor Neural Network classifier is applied for extracting images to enhance the
classifier performance using geometric functions. Four metrics are computed to evaluate the proposed IGGL-GNNC
method namely, sensitivity, specificity, prediction accuracy and PSNR for different plant leaf images acquired from
Plant Village Image dataset. Results indicate that proposed method can effectively detect plant leaf disease in plant
leaf images.