MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641076797 A) filed by Vellore Institute Of Technology on June 21, 2026, for Explainable Multi-Scale Few-Shot Learning System For Hydroponic Nutrient Deficiency Detection.
Inventors include Manikandan G; Pokuri Venkata Sri Vishal; and Harshvardhan Sinha.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: The present invention relates to an explainable multi-scale few-shot learning system (100) for nutrient deficiency classification from plant leaf images. The system (100) comprises a data acquisition module configured to receive leaf images, a dataset processing and conversion module configured to preprocess image data, a few-shot learning framework configured to generate episodic learning tasks, a multi-scale embedding module configured to extract feature representations from multiple neural-network layers and generate embedding vectors, and a prototype computation and classification module configured to generate class prototypes and classify query images based on similarity evaluation. The system (100) further comprises an explainability module configured to retrieve reference samples and generate similarity-based explanation information, and a stability-aware decision module configured to evaluate prediction consistency and generate stability information. The invention provides an integrated architecture combining multi-scale feature learning, prototype-based classification, explanation generation, and stability assessment.
Disclaimer: Curated by HT Syndication.
