Banana Leaf Disease Classification Using SqueezeNet, AlexNet and MobileNet

Abstract

Banana production is a vital component of global agriculture, facing significant challenges due to various leaf diseases. These diseases can cause substantial yield losses, impacting both farmer livelihoods and industry development. Early and accurate disease detection is crucial for implementing effective management strategies. This study explores the application of Convolutional Neural Network (CNN) for banana leaf disease classification. Three pre-trained CNN architectures, SqueezeNet, AlexNet, and MobileNet, were evaluated for their ability to distinguish between Black Sigatoka, Fusarium Wilt, and healthy banana leaves. A comprehensive dataset containing 3000 images per class was employed for training and testing, with an 8:1:1 train-validation-test split. Performance evaluation metrics included accuracy and loss rate. Among the evaluated models, MobileNet achieved the highest accuracy (94.89%) and the lowest loss (0.1484), demonstrating its effectiveness in banana leaf disease detection. These findings suggest the potential of CNNs as a valuable tool for precision agriculture applications.