MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641123522 A) filed by M. Gunasundhari; B. Kavenilaa; J. Jenova; A. Kaviya; and V. Jaihindh on October 15, 2026, for Emotional Recognition Using Multimodel Llm.
Inventors include M. Gunasundhari; B. Kavenilaa; J. Jenova; A. Kaviya; and V. Jaihindh.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: Emotional Recognition using Multimodal LLM is an AI-based project that identifies a person's emotions by analyzing text, speech, and facial expressions. The system processes each input separately using machine learning and deep learning models, then combines the results with a Large Language Model (LLM) to predict the final emotion more accurately. The project is developed using Python, OpenCV, DeepFace, Librosa, BERT, and Stream lit. It can recognize emotions such as Happy, Sad, Angry, Fear, Surprise, and Neutral. This project demonstrates how multiple input sources can improve the accuracy and reliability of emotion recognition. The proposed invention, "Emotional Recognition using Multimodal LLM," is an Al-based system that detects human emotions by analyzing text, speech, and facial expressions. The system processes each input separately and combines the results using a Large Language Model (LLM) to provide accurate emotion recognition. It can identify emotions such as Happy, Sad, Angry, Fear, Surprise, and Neutral. This invention enhances emotion detection accuracy and can be used in applications such as healthcare, education, customer support, and intelligent human-computer interaction.
Disclaimer: Curated by HT Syndication.
