Plant Species in Classification using Convolutional Neural Networks on Leaf Images

Recently, a classification of plant species has a limited method by dealing with large data sets in term of plant recognitions.This study investigates the application of convolutional neural networks (CNNs) for plant species classification using leaf images.Recognizing plant species accurately is an important task in botany, forestry, and environmental monitoring, and leveraging deep learning techniques offers significant potential in automating this process.The research utilizes a subset of the Leafsnap dataset, which includes labeled leaf images of various plant species native to the North Eastern United States.Transfer learning techniques are employed to evaluate and compare the performance of three CNN models: the traditional AlexNet, and two state-of-the-art models, MobileNetv4 and LMFRNet.The primary objective is to develop a lightweight yet effective model for leaf classification.Experimental results show that the recently proposed CNN models achieved classification accuracies exceeding 95% on the selected dataset, suggesting the effectiveness of modern CNNs for plant species recognition tasks.