MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641100713 A) filed by Jeppiaar Institute Of Technology on August 20, 2026, for "herbsathi" - An Ai Voice-Based Herbal Assistant For Health-Personalized Medicinal Leaf Identificati.
Inventors include Ashika. J; Mounika. V; Ruththika S; Dr. J. Venu Gopala Krishnan; Dr. T. Sripriya; Dr. W. Nancy; Dr. M. Benisha; M. Indumathi; Bharathi M; Karthika. T; M. Nithya; S. Revathy; and M. Sudhagar.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: This patent disclosure covers a novel web-based system for automated medicinal herbal leaf identification and personalized health-based recommendation. Traditional herbal medicine practice largely relies on manual identification by experts, which is time-consuming, inaccessible to common users, and prone to error, often leading to misuse of herbal remedies without regard to an individual's health condition. This disclosure proposes a system wherein the user uploads or scans an image of a plant leaf through a web interface, and the system first evaluates the quality of the uploaded image before further processing to ensure reliable identification. The system employs image processing techniques including Gaussian blur, HSV-based segmentation, and color, shape, texture and edge feature extraction, coupled with a machine learning-based classification model, to identify the plant species from the captured leaf image. Upon identification, the system retrieves and displays the medicinal properties of the identified plant, along with specific guidance on suitability of use based on individual health parameters such as age, existing health conditions, allergies, and pregnancy status, thereby flagging herbs that may be unsuitable for a given user. The system further integrates HerbSathi, an Al voice-based herbal assistant that interacts, with the user through spoken conversation - converting the user's speech to text, processing the query, generating an appropriate response, and converting that response back to speech - so that the user can ask questions and receive guidance about the identified plant hands-free and in natural language, without needing to type. This voice interface also unobtrusively collects the user's health information needed for personalization, making the platform accessible even to users with limited literacy or typing ability. To assist users unfamiliar with herbal preparation methods, the system provides instructional preparation videos corresponding to the identified plant. Additionally, the system hosts a reference library comprising images, medicinal uses, and safety information of multiple herbal plants for general learning purposes. The proposed system thereby provides an accessible, personalized, and safety-conscious platform for herbal plant identification and utilization, combining image processing, machine learning, and an Al voice assistant in a single integrated web application
Disclaimer: Curated by HT Syndication.
