Voice Data Fragmentation for Privacy Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice processing systems face challenges in protecting user privacy as voice samples, when transmitted for processing, can expose sensitive information, compromising user privacy.
Innovation Solution
A system that captures voice data locally, fragments it into scrambled data fragments, and transmits these for analysis, minimizing exposure of privacy-sensitive information by reordering the fragments based on predefined parameters, ensuring that privacy is maintained during voice-related service processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If voice samples are transmitted to external systems for processing, then voice-related services can be provided, but user privacy-sensitive information is exposed and compromised
Solution Approach 1:
The voice data is divided into multiple fragments before transmission. Each fragment contains only a portion of the original voice information, making it impossible to reconstruct meaningful voice data or extract privacy-sensitive information from individual fragments. This segmentation approach enables external processing while protecting user privacy.
Solution Approach 2:
A fragment identification value is introduced as an intermediary element that links voice fragments to their original source without containing the actual voice content. This intermediary enables the external system to process and match fragments appropriately while the privacy-sensitive voice information remains protected and cannot be reconstructed from the fragments alone.
2Object-affected harmful factors
If voice data is fragmented and scrambled before transmission, then user privacy is protected, but the complexity of the processing system increases
Solution Approach 1:
The voice data fragmentation, scrambling, and fragment identification value assignment are performed in advance before transmission to the external system. This preliminary processing eliminates the need for complex real-time decryption or reconstruction algorithms in the external system, reducing overall system complexity while maintaining privacy protection.
Solution Approach 2:
The system transforms the voice data from its original continuous form into discrete fragments with associated identification values. This parameter transformation simplifies the data structure for transmission and processing, enabling the external system to handle fragmented data more efficiently without requiring complex privacy-preserving computation infrastructure.
3Object-affected harmful factors
If voice samples are processed locally without transmission, then user privacy is maintained, but voice-related services cannot be provided
Solution Approach 1:
By segmenting voice data into non-reconstructible fragments, the system enables external processing infrastructure to be utilized for providing voice services while ensuring that no privacy-sensitive information is exposed during transmission or processing. Each fragment alone is insufficient for service reconstruction, maintaining privacy while enabling service capability.
Solution Approach 2:
The fragment identification value serves as a mediator that allows the external system to perform matching and processing operations without accessing the actual voice content. This intermediary mechanism bridges the gap between local privacy preservation and remote service delivery, enabling voice-related services to function while maintaining user anonymity.
Data Source
AI summary
Technologies for privately processing voice data include a compute device configured to continually or periodically capture voice data of a user by the compute device. The captured voice data is processed to remove or reduce the user's privacy-sensitive information. For example, the compute device fragments the captured voice data to generate a set of voice data fragments and further scrambles the voice data fragments to generate scrambled voice data fragments having a sequential order different from the plurality of voice data fragments. To scramble the voice data fragments, the compute device may reorder the voice data fragments such that each fragment is repositioned from its corresponding original sequential position in the voice data by a particular number of words, syllables, or phrases.