Secure Audio Recognition Using Isolated Memory and Encryption

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing audio security mechanisms for live audio captured via microphones are insecure and prone to attacks, as they lack the robust protection provided for pre-recorded audio.

Innovation Solution

A system for secure end-to-end audio recognition is implemented, using a client computing device and a server that exchange cryptographic keys to establish a secure connection, with a security engine encrypting and decrypting audio data in a secure memory, isolating it from the central processing unit to prevent unauthorized access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If audio digital rights management (DRM) systems and hardware reinforced audio protection mechanisms such as Protected Audio Video Path (PAVP) are used, then pre-recorded audio is secured, but live audio captured via microphone remains insecure and prone to attack

Engineering Contradiction:

Improveaudio securityVSAvoidcoverage of audio types

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent extends audio protection mechanisms to cover both pre-recorded audio and live audio captured via microphone, creating a universal security system that handles multiple audio types through a unified approach involving secure memory, encryption, and isolated processing paths

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Ease of operation

If audio data is processed through standard computing systems, then audio recognition functionality is achieved, but the audio data becomes vulnerable to interception and malware

Engineering Contradiction:

Improveaudio processing capabilityVSAvoidinterception and malware vulnerability

Core Design Contradiction:

Ease of operationVSObject-affected harmful factors

Solution Approach 1:

The patent segments the audio processing system into isolated components: secure memory for audio data storage, dedicated security engines for encryption/decryption, and isolated audio pipelines that prevent unauthorized access while maintaining processing functionality

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces secure memory and security engines as intermediary layers between the microphone/audio input and the processing systems, creating protected pathways that enable audio processing while blocking interception and malware attacks

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS9577784B2System, device, and method for securing voice authentication and end-to-end speech interaction

Publication Date: 2017.02.21 INTEL CORP

AI summary

A method, device, and system for secure end-to-end audio recognition is disclosed. A client device launches an application that connects with a server. The client device and server exchange cryptographic keys and establish a secure connection and a shared cryptographic key. The server transmits an encrypted audio prompt to the client device. The client device decrypts the encrypted audio prompt and stores the decrypted audio prompt in secure memory inaccessible to the operating system using an audio engine of the client device. The audio engine then retrieves the audio and renders it for the user through the speakers of the client device. The client device captures the user's audio response with a microphone and stores the audio response in the secure memory. The stored audio response is encrypted and transmitted to the server.