Call Handling System Using Speech Analysis for Spoofed Number Fraud
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to effectively classify and handle unwanted telephone calls, particularly those using spoofed numbers, which can lead to fraud and disrupt legitimate communications.
Innovation Solution
A system and method that intercepts calls, generates a call recording, determines attributes using an automatic speech recognizer, and classifies calls as unwanted using a trained classifier, allowing for early termination and secure handling of such calls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If call blocking based on spoofed numbers is implemented, then fraud prevention is improved, but legitimate calls from banks and institutions are blocked
Solution Approach 1:
The system segments the call classification process into multiple stages: initial filtering based on caller ID, speech recognition analysis of call content, and classifier-based decision making. This allows nuanced differentiation between fraudulent and legitimate calls using spoofed numbers, preventing blanket blocking of institutional calls while catching fraud.
Solution Approach 2:
The system performs preliminary speech recognition and attribute analysis during the call setup phase before the call is fully established. By analyzing call attributes, speech patterns, and content in advance, the system can identify fraudulent calls early and terminate them before they cause harm, while allowing legitimate institutional calls to proceed.
2Measurement precision
If comprehensive call analysis is performed, then call classification accuracy is improved, but processing time increases
Solution Approach 1:
The system performs speech recognition and attribute extraction during call setup and in parallel with initial processing, rather than sequentially after the call is fully established. This preliminary analysis allows accurate classification without significantly delaying call processing for legitimate calls.
Solution Approach 2:
For calls that pass initial filtering criteria (e.g., recognized institutional numbers, clear speech patterns), the system can skip extensive analysis and allow the call to proceed quickly. Comprehensive analysis is applied selectively to suspicious calls, reducing overall processing time while maintaining accuracy for fraud detection.
3Object-affected harmful factors
If early termination of unwanted calls is implemented, then fraud damage is reduced, but false positives disrupt legitimate communications
Solution Approach 1:
The termination decision is segmented into multiple verification stages: initial risk assessment, speech content analysis, attribute validation, and final classifier decision. Calls must fail multiple checks before termination, reducing false positives while maintaining early termination capability for confirmed fraud.
Solution Approach 2:
The system incorporates feedback loops where classification results are continuously refined based on outcomes. False positives are fed back into the training data to improve the classifier, and user feedback on misclassified calls allows the system to learn and reduce future false terminations of legitimate calls.
Data Source
AI summary
Disclosed herein are systems and methods for handling unwanted telephone calls. In one aspect, an exemplary method comprises, intercepting a call request for a call from a terminal device of a calling party to a terminal device of a called party, generating a call recording containing media data transmitted within a connection established by the intercepted call request, determining attributes of the generated call recording, classifying the call as an unwanted call based on the determined attributes, wherein the classification is performed by a classifier trained on previously collected unwanted calls, and wherein the call is classified as unwanted when the attributes belong to an unwanted call class that is known, and handling the call in accordance with the classification of the call, the handling including at least securing information of the call.