MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202641123018 A) filed by Mlr Institute Of Technology; and Marri Laxman Reddy Institute Of Technology And Management on October 14, 2026, for A Cognitive Multi-Layer Acoustic Fingerprinting System For Forensic Speaker Identification And Voice Integrity Assessment.
Inventors include Dr. Venkata Subba Reddy Gade; Dr. P. Yakaiah; Mr. Ch. Vijay; and Mr. Kshamith Reddy.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: The present invention relates to a computer-implemented acoustic analysis system for forensic speaker identification and voice integrity assessment. The system comprises a signal acquisition module configured to receive questioned and reference speech samples from microphones, telecommunication channels, surveillance recordings, digital audio archives, and forensic databases; a preprocessing module configured to perform filtering, voice activity detection, noise suppression, channel equalization, frame segmentation, silence trimming, and gain normalization; and a hierarchical acoustic fingerprint extraction engine configured to derive physiological, behavioural, micro-acoustic, and authenticity-related features from the speech samples. A feature fusion and inference engine is configured to assign weights to extracted features, generate discriminative representations, compute similarity measures, and evaluate probabilistic hypotheses for speaker comparison. A voice integrity verification module analyses spectral anomalies, phase discontinuities, harmonic inconsistencies, and modulation irregularities to identify potential replay attacks, recording tampering, voice conversion, and synthetic speech generation, and generates a voice integrity score. An explainable forensic decision module produces speaker similarity scores, confidence estimates, feature contribution rankings, authenticity assessment results, and audit records for review by forensic experts. The system integrates acoustic preprocessing, multi-layer fingerprint extraction, speaker comparison, voice integrity verification, and explainable decision support within a unified computational architecture for analysing speech recordings under varying noise, compression, language, and recording-channel conditions. 5 Claims & 3 Figure.
Disclaimer: Curated by HT Syndication.
