MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202621094610 A) filed by Symbiosis International Deemed University on August 04, 2026, for Explainable Ai-Based System And Method For Multi-Class Tomato Leaf Disease Detection.
Inventors include Prof. Riddhi Doshi; Ojaswini Thote; Adhiraj Kulkarni; and Tanay Soman.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: ABSTRACT EXPLAINABLE AI-BASED SYSTEM AND METHOD FOR MULTI-CLASS TOMATO LEAF DISEASE DETECTION The present invention provides an explainable AI-based system (100) and method for multi-class tomato leaf disease detection. The system (100) comprises an image acquisition module (110), an image preprocessing module (120), a data augmentation module (130), a transfer learning-based classification module (140) comprising a pre-trained MobileNetV2 backbone (142) with frozen convolutional layers and a custom classification head (144) configured to classify tomato leaf images into ten categories comprising nine diseases and one healthy class, and an explainability module (150) comprising a Grad-CAM visualization unit (152) for generating gradient-weighted class activation heatmaps and a LIME interpretation unit (154) for generating superpixel-based local interpretable explanations. The system (100) outputs a disease class label (162), confidence score (164), heatmap visualization (166) and superpixel explanation (168) through an output and visualization module (160), achieving 95.86 percent accuracy with interpretable diagnostic outputs. [
Disclaimer: Curated by HT Syndication.
