MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112717 A) filed by Dr. Parkavi A; Uzma Sulthana; Dr. Krupa Rasane; Dr. Manjula L; Dr. Devaraju B M; Dr. Shaleen Bhatnagar; Nikhil Y N; Sanjay J; Utkarsh Patel; and Shashwat Kumar on September 20, 2026, for Hardware-Accelerated Educational Multi-Speaker Audio Analysis System.
Inventors include Dr. Parkavi A; Uzma Sulthana; Dr. Krupa Rasane; Dr. Manjula L; Dr. Devaraju B M; Dr. Shaleen Bhatnagar; Nikhil Y N; Sanjay J; Utkarsh Patel; and Shashwat Kumar.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: The present invention relates to a hardware-accelerated educational multi-speaker audio analysis system comprising an audio capture / media input interface (10), preprocessing module (20), speech activity detection module (30), speaker diarization module (40), transcription module (50), transcript refinement module (60), structured intermediate representation layer (70), acoustic feature extraction module (80), semantic embedding module (90), temporal interaction modelling module (100), knowledge graph construction module (110), centrality analysis module (120), projection module (130), and visualization dashboard (140). A multi-core CPU (150), GPU accelerator (160), memory (170), SSD storage (180), display interface (190), and network interface (200) support hardware-based processing. The system receives classroom audio or video, generates speaker-labelled timestamped transcripts, extracts acoustic and semantic features, computes semantic velocity, prosodic weight, interaction energy, and multimodal prosodic-semantic score. The system further constructs a weighted conversational knowledge graph and renders discourse pulse, semantic trajectory, speaker ranking, concept ranking, transcripts, reports, and data exports. (Accompanied Figure No. 1-3)
Disclaimer: Curated by HT Syndication.
