MUMBAI, India, Sept. 7 -- Intellectual Property India has published a patent application (202641104397 A) filed by Srm Trp Engineering College on August 31, 2026, for Real-Time Emotion-Aware Virtual Interviewer Using Multi-Modal Deep Learning.
Inventors include Mrs. S. Devapriya; Seshavahini G; Sriram R; and Surya Prakasam V.
The application for the patent was published on September 04, 2026, under issue no. 36/2026.
Abstract: This invention presents a Real-Time Emotion-Aware Virtual Interviewer using Multi-Modal Deep Learning, designed to enhance the efficiency and fairness of interview processes through intelligent automation. The system captures candidate responses using multiple input modalities including video, audio, and text, enabling comprehensive analysis of emotional and behavioral patterns. Facial expressions are analyzed using Convolutional Neural Networks (CNN) to detect emotions such as confidence, stress, and nervousness. Speech signals are processed using acoustic feature extraction techniques to evaluate tone, pitch, and speaking patterns. Text responses are analyzed using Natural Language Processing (NLP) to assess sentiment, coherence, and response quality. The system integrates these modalities to perform real-time emotion detection and generates a unified emotional profile of the candidate. Based on this analysis, the virtual interviewer dynamically adapts the questioning strategy, making the interaction more personalized and context-aware. Additionally, the system evaluates overall candidate performance by analyzing communication effectiveness, emotional stability, and response accuracy. It generates automated feedback reports with actionable insights for improvement.
Disclaimer: Curated by HT Syndication.
