MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641076836 A) filed by Mr. Dhruva M S; and Dr. Sunitha R on June 22, 2026, for A Cross Modal Emotion Recognition Framework For Deep Fusion Of Visual And Acoustic Cues Using Deep Learning Techniques To Enhance Accuracy.
Inventors include Mr. Dhruva M S; and Dr. Sunitha R.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: The present invention relates to the development of a framework for cross-modal emotion recognition is presented that uses deep fusion of visual and acoustic features with the help of advanced deep learning approaches to enhance the performance of emotion classification. This framework collects facial expressions, head motions, speech, and acoustic features from different sources, which are processed by specialized feature extraction components. Visual features are extracted by employing convolutional neural networks, whereas acoustic features are extracted by employing both recurrent neural networks and transformers, which are able to capture temporal correlations in the data. These features are fused using a cross-modal fusion component that utilizes attention and adaptive weightings to create an emotional representation. In addition, the framework utilizes a classification component for classifying emotions into categories including happiness, sadness, anger, fear, surprise, and neutral emotions. FIG.1
Disclaimer: Curated by HT Syndication.
