Impact of Hyperparameter Tuning for Identification and Classification of Plant Leaf Diseases: A Deep Learning Approach

Abstract

Agriculture is a prominent sector that contributes significantly to the country's economic development, accounting for 20.19% of gross domestic product (GDP) as of the year 2020–2021. Technologies like Internet of Things, Machine Learning (ML), Deep Learning (DL), and Artificial Neural Networks (ANN) provide the most effective and feasible solutions. This aids in making different domain modernization through automation in agricultural fields with minimal human intervention. This paper presents a convolutional neural network framework using the PlantVillage dataset for tomato plants affected by several diseases. With rigorous experimentation and parameter tuning the impact of hyperparameter on the model, performance is observed and the best fit model is considered for the experimentation.