Deep Learning Perspectives for Tulsi Leaf Detection and Disease Diagnosis using CNN and ResNet Models

Tulsi is a widely used medicinal plant known for its numerous health benefits, but diseases affecting its leaves can reduce plant growth, quality, and medicinal value. Early disease detection is important for effective plant management. however, identifying leaf diseases often requires expert knowledge that may not always be available. This work presents a deep learning–based system for automated Tulsi leaf disease diagnosis using image analysis. The proposed system operates in two stages. First, a Convolutional Neural Network (CNN) validates whether the uploaded image contains a plant leaf. Second, a ResNet50 model based on transfer learning classifies the leaf image into one of five categories: Bacterial Wilt, Healthy, Leaf Spot, Powdery Mildew, and Web Blight. Image preprocessing techniques are applied to improve classification performance. In addition to disease detection, the system provides disease information, symptoms, and treatment recommendations to assist users in managing plant health. The proposed framework offers a simple, accurate, and user-friendly solution that can support farmers, researchers, students, and home gardeners in the early diagnosis of Tulsi leaf diseases.