Cassava Leaf Disease Classification with Deep Learning

Abstract

Inspired by the current trend in artificial intelligence and neural network research, and with the increase in accuracy of models with deep neural network architecture (DNN), our work is dedicated to developing and training a deep neural network to extract meaningful patterns of diseases from leaves. We show how DNNs are applied to classification problems – classifying cassava leaf disease with a form of advanced convolutional neural networks with compound scaling (Efficient Nets) using supervised learning approach.