오세희, 오준형, 이성민, 이해진, 이효민,

Vol. 35, No. 1, pp. 157-165, 2월. 2025
10.13089/JKIISC.2025.35.1.157, Full Text:
Keywords: Privacy-Protection, information security, Voice anonymization, Unstructured Data, Re-identification
Abstract

The rise of big data has underscored the need for data protection. Voice data, a form of unstructured data, is particularly sensitive as it can reveal personal identities on its own or in combination with other data, heightening speaker identification risks. To address this, anonymizing voice data has become essential. However, even anonymized speech may be vulnerable to re-identification under certain conditions. This study proposes a framework for evaluating the re-identification risk of anonymized speech and determining optimal anonymization parameters. This approach seeks to balance privacy protection and data utility, providing guidelines for the safe use of voice data in the information society.

Statistics
Cite this article