Herbal Plants Leaf Image Classification Using Deep Learning Models Based On Augmentation Approach

Abstract

As the world grows daily, people are shifting towards renewable energy sources and natural resources for healthcare as remedies. Herbs are the natural source of medicine in place of synthetic drugs. Herbs are those kinds of plants used to cure the disease of humans or animals. Identification is the most critical aspect of using herbs as medicines. The problem of identification of these herbal plants is trying to be solved using Deep Learning (DL) models in this paper. Here the authors have successfully classified 25 categories of herbal plant leaf images using different deep learning models with the highest test accuracy of 97.68% on original data and 98.08% on augmented data. This research work is an extension of the previous research work by the same authors, titled "Herbal plants leaf image classification using machine learning approach", in which the six classical Machine Learning (ML) algorithms have been applied to original data and the highest classification accuracies of 82.51% has been achieved by using Multi-layer Perceptron (MLP) classifier.

Keywords: Computer vision, Deep learning, Herbal plants, Image classification, Image processing, Machine learning, RGB image

JEL Classification: C02