Voice Authentication for Identity Verification With Coercion Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional user verification techniques are susceptible to counterfeiting and fail to determine whether biometric data is coerced or provided by an intellectually incapacitated individual, lacking robustness in ensuring genuine and voluntary consent.

Innovation Solution

A system utilizing voice analysis, geolocation, and consent verification components to authenticate user identity, ensuring voluntary and competent consent through voice recordings, geolocation, and analysis of physiological indicators like nystagmus and pupil size, with encryption for security.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If conventional biometric verification techniques (facial recognition, fingerprint recognition) are used, then verification can be performed without documents, but the system cannot determine whether the biometric data is a copy (mask, 3D-printed mold) or coerced

Engineering Contradiction:
Improvedocument-free verificationVSAvoiddetection of copies and coercion
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The verification process is divided into multiple independent analysis components: voiceprint analysis for identity verification, liveness detection to identify copies, and physiological state analysis to detect coercion. Each component analyzes a different aspect of the biometric data independently, allowing the system to comprehensively assess whether the verification is genuine without relying on a single method.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The voice-based verification system serves multiple functions simultaneously: it verifies identity through voiceprint matching, detects liveness to identify masks or 3D-printed fingerprints, and assesses physiological state to detect coercion or intellectual incapacitation. This multi-functional approach is achieved through analyzing various acoustic characteristics and physiological indicators within the voice data.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Reliability

If voice analysis is added to detect liveness and physiological state, then detection of copies and coercion is improved, but system complexity increases

Engineering Contradiction:
Improvedetection of copies and coercionVSAvoidverification system structure
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system combines voiceprint analysis, liveness detection, and physiological state analysis into a single integrated verification process. Instead of using separate systems for each function, the patent merges these capabilities into one voice-based verification module that processes all types of data simultaneously, reducing overall system complexity while maintaining comprehensive verification capabilities.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

A single voice analysis system performs multiple verification functions: identity verification through voiceprint matching, liveness detection to identify copies, and physiological state analysis to detect coercion. By making the voice analysis system multi-functional, the patent avoids the need for separate dedicated systems for each function, thereby reducing overall device complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250288232A1Methods and Systems for Identity Verification Using Voice Authentication
Publication Date: 2025.09.18 AEGIS-CC LLC
  • US20250288232A1 patent drawing
  • US20250288232A1 patent drawing
  • US20250288232A1 patent drawing

AI summary

Aspects of the disclosure relate to processing images and identifying features in the images. Images of a user captured using a camera are processed to enhance subsequent analysis. A face is located in the images using Haar cascades, a Histogram of Oriented Gradients, a Viola-Jones framework, and/or a first deep learning algorithm. An eye in the face is located using a plurality of located facial landmarks and/or a convolutional neural network. Movements of the eye in the images is determined. Eye jerking in the images is detected using the determined eye movements. A distance of the detected eye jerking of the eye and how long the detected eye jerking lasted is determined and used to generate an indicator. At least partly in response to the indicator, messages are generated and transmitted to electronic destinations.