MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641079049 A) filed by Sr University on June 25, 2026, for A Multimodal, Self-Supervised, Federated And Explainable Framework For Crop Disease Detection With Edge-Optimized Deployment.
Inventors include Lalitha Manglaram; and Dr. Sudersan Behera.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: A MULTIMODAL, SELF-SUPERVISED, FEDERATED AND EXPLAINABLE FRAMEWORK FOR CROP DISEASE DETECTION WITH EDGE-OPTIMIZED DEPLOYMENT The present invention relates to a multimodal, self-supervised, federated and explainable framework for crop disease detection with edge-optimized deployment. The framework comprises a self-supervised representation learning module, a federated learning module, a multimodal fusion engine, an explainability module, and an edge deployment and optimization unit. The system learns disease-related features from unlabeled agricultural data using self-supervised learning techniques and performs privacy-preserving federated training without transferring raw farm data. Multimodal inputs including crop images, soil characteristics, environmental measurements, geospatial metadata, and temporal information are fused using advanced neural architectures to improve disease diagnosis accuracy. The explainability module generates interpretable visual and textual outputs identifying factors influencing disease predictions. Edge optimization techniques including quantization, pruning, lightweight model architectures, and on-device caching enable real-time disease detection on low-power devices. The invention provides an accurate, privacy-preserving, explainable, and scalable solution for crop disease diagnosis in resource-constrained agricultural environments.
Disclaimer: Curated by HT Syndication.
