Abstract
This paper aims to frame a new rice disease prediction model that included three major phases. Initially, median filtering (MF) is deployed during pre-processing and then ‘proposed Fuzzy Means Clustering (FCM) based segmentation’ is done. Following that, ‘Discrete Wavelet Transform (DWT), Scale-Invariant Feature Transform (SIFT) and low-level features (colour and shape), Proposed local Binary Pattern (LBP) based features’ are extracted that are classified via ‘Multi-Layer Perceptron (MLP) and Long Short Term Memory (LSTM)’ and predicted outcomes are obtained. For exact prediction, this work intends to optimise the weights of LSTM using Inertia Weighted Salp Swarm Optimisation (IW-SSO) model. Eventually, the development of IW-SSO method is established on varied metrics.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
Data sharing is not applicable to this article as no new data were created or analysed in this study.
Additional information
Notes on contributors
Sakhamuri Sridevi
Sakhamuri graduate of MCA, MTech, currently pursuing part-time Ph.D. in CSE from KL University. She is working as an assistant professor in the Department of ECM. Interested areas of research are data mining, and machine learning in the field of Agriculture. She is an expert in Software Engineering, Web technologies, Discrete Mathematics, Object Oriented Programming. Done certifications in Java and Python.
K. Kiran Kumar
K. Kiran Kumar is currently working at Koneru Lakshmaiah Education Foundation Vaddeswaram, Guntur Dist, AP, India.