system
The voice-based fraud detection system addresses the inadequacies of static fraud detection by integrating speech recognition, machine learning, and emotional analysis to provide real-time, comprehensive fraud alerts, ensuring elderly individuals are protected from evolving scams.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- SOFTBANK GROUP CORP
- Filing Date
- 2024-12-09
- Publication Date
- 2026-06-19
AI Technical Summary
Existing fraud detection systems are inadequate in accurately and timely identifying evolving fraud methods targeting the elderly, particularly through voice-based scams, as they rely on static analysis and fail to consider emotional nuances and real-time adaptability.
A voice-based fraud detection system that utilizes a high-performance microphone to capture ambient sound, converts it into text data using speech recognition, analyzes it for fraud patterns with machine learning algorithms, and sends alerts to registered contacts, while incorporating an emotion engine to assess emotional states and continuously update fraud databases.
The system provides real-time, accurate fraud detection by analyzing both voice content and emotional cues, reducing the risk of elderly individuals falling victim to scams and ensuring the database remains updated with the latest fraud patterns.
Smart Images
Figure 2026100717000001_ABST