Hybrid AI Model DLFL-TLDNet Classifies 10 Tomato Leaf Diseases with 97.91% Accuracy

A hybrid artificial intelligence framework called DLFL-TLDNet was released on Sunday combining deep learning and fuzzy logic to classify tomato leaf diseases and estimate their severity with 97.91 percent accuracy. Developed by Md Hasibul Islam and Fernaz Narin Nur in research published in Scientific Reports, the system addresses a critical gap in agricultural management, where timely and precise identification supports sustainable farming and food security.

While numerous disease detection models exist in the current literature, existing approaches frequently lack severity estimation capabilities. To overcome this limitation, the DLFL-TLDNet architecture integrates a pre-trained InceptionV3 convolutional neural network to categorize ten distinct tomato leaf disease categories. According to comparative experimental evaluations detailed by Scienmag, this configuration outperformed several alternative baseline deep learning architectures, specifically DenseNet121, VGG16, ResNet152, EfficientNet-B0, and MobileNetV2.

To quantify disease progression, the framework employs a fuzzy logic-based inference system driven by disease-specific segmentation data. This inference system evaluates two primary physical parameters derived from segmented images: the total lesion area and the lesion count. By combining these geometric measurements within a fuzzy logic structure, the system calculates clear severity scores categorized into Primary, Medium, and Extreme levels.

Transparency and user trust are enhanced through explainable artificial intelligence techniques, incorporating Gradient-weighted Class Activation Mapping (Grad-CAM) visualizations and color-coded overlays. To establish practical utility for precision agriculture and early, sustainable tomato disease management, the authors implemented a prototype web application. Through this web application interface, users can upload leaf images for real-time diagnosis and automated severity reporting to obtain immediate diagnostic feedback directly from the system.

Photo of author

Health Desk

Health Desk at Archyde.

New Book Chronicles 200 Years of Italian Migration in Thurgau