Cassava Leaf Disease Detection Using Deep Learning with ResNet-50, VGG-16, and EfficientNet Architectures

Cassava is a vital crop in many tropical regions, including Bandar Lampung, Indonesia, where it serves as both a staple food and a nutritional resource. However, cassava production is severely threatened by various leaf diseases such as Cassava Brown Streak Disease (CBSD), Cassava Green Mite (CGM), Cassava Bacterial Blight (CBB), and Cassava Mosaic Disease (CMD), which significantly reduce crop yield and quality. Traditional disease diagnosis methods are often costly and inaccessible to many farmers, leading to delayed interventions and losses. To address this issue, a deep learning-based approach is proposed for automated cassava leaf disease detection using Convolutional Neural Networks (CNNs), specifically ResNet-50, VGG-16, EfficientNet, and EfficientNetB3 architectures. These models are capable of extracting complex image features to achieve high classification accuracy. The proposed system offers an early detection mechanism that is cost-effective, scalable, and easy to use, ultimately promoting sustainable agricultural practices.