Intellectual Property India Publishes Patent Application for Ai-Based Cross-Modal Fraud Detection System For Identifying Digital Arrest Scams Through Voice Calls And Forged Digital Evidence

MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641124637 A) filed by Mr. Siva Yenikepalli; Dr. Devinder Singh Anand; Kalaiyarasi R; B. Soujania; Dr. K. Muthumanickam; Dr. Rajnish Bishnoi; Dr. Dharmendra Kumar Kumawat; A. Sabari Gayathri; Dr. Priyanka; Dr. S. Sri Jamiya; Dr Sumanta Bhattacharya; and Dhivya S on October 17, 2026, for Ai-Based Cross-Modal Fraud Detection System For Identifying Digital Arrest Scams Through Voice Calls And Forged Digital Evidence.

Inventors include Mr. Siva Yenikepalli; Dr. Devinder Singh Anand; Kalaiyarasi R; B. Soujania; Dr. K. Muthumanickam; Dr. Rajnish Bishnoi; Dr. Dharmendra Kumar Kumawat; A. Sabari Gayathri; Dr. Priyanka; Dr. S. Sri Jamiya; Dr Sumanta Bhattacharya; and Dhivya S.

The application for the patent was published on October 30, 2026, under issue no. 44/2026.

Abstract: AI-Based Cross-Modal Fraud Detection System for Identifying Digital Arrest Scams through Voice Calls and Forged Digital Evidence is the proposed invention. The proposed invention relates to an artificial intelligence-based cross-modal fraud detection system for identifying digital arrest scams through integrated analysis of voice calls and associated digital evidence. The system comprises an audio forensic analysis module, a digital evidence inspection module, a semantic extraction module, a multimodal representation encoder, a cross-modal evidence correlation engine, and an uncertainty-aware risk assessment module. Self-supervised audio representation learning is employed to identify suspicious speech characteristics and potential synthetic voice artifacts, while a vision-language model analyses purported official documents, identity credentials, and manipulated screenshots. An uncertainty-aware cross-attention fusion mechanism correlates acoustic, textual, visual, and contextual features to identify inconsistencies between spoken claims and submitted evidence. The system generates an explainable fraud risk score and provides contextual warnings, independent verification recommendations, and structured incident reports. The invention facilitates integrated assessment of impersonation attempts, improves detection of coordinated voice-and-document deception, and supports privacy-conscious digital fraud prevention through controlled evidence processing and human-assisted verification.

Disclaimer: Curated by HT Syndication.