Utilizing Speaker Profiles for Impersonation Audio Detection
Abstract: Fake audio detection is an emerging active topic. A growing number of literatures have aimed to detect fake utterance, which are mostly generated by Text-to-speech (TTS) or voice conversion (VC). However, countermeasures against impersonation remains an underexplored area. Impersonation is a fake type that involves an imitator replicating specific traits and speech style of a target speaker. Unlike TTS and VC, which often leave digital traces or signal artifacts, impersonation involves live human beings producing entirely natural speech, rendering the detection of impersonation audio a challenging task. Thus, we propose a novel method that integrates speaker profiles into the process of impersonation audio detection. Speaker profiles are inherent characteristics that are challenging for impersonators to mimic accurately, such as speaker's age, job. We aim to leverage these features to extract discriminative information for detecting impersonation audio. Moreover, there is no large impersonated speech corpora available for quantitative study of impersonation impacts. To address this gap, we further design the first large-scale, diverse-speaker Chinese impersonation dataset, named ImPersonation Audio Detection (IPAD), to advance the community's research on impersonation audio detection. We evaluate several existing fake audio detection methods on our proposed dataset IPAD, demonstrating its necessity and the challenges. Additionally, our findings reveal that incorporating speaker profiles can significantly enhance the model's performance in detecting impersonation audio.
Primary Subject Area: [Content] Media Interpretation
Secondary Subject Area: [Experience] Multimedia Applications
Relevance To Conference: Given the devastating consequences of fake audio on both individuals and society, fake audio detection has become an urgent and essential task that needs to be addressed. However, most previous fake audio detection research has primarily focused on four kinds of fake types: text-to-speech, voice conversion, replay and partially fake. The domain of voice spoofing through impersonation remains relatively underexplore. Existing mainstream studies on fake audio detection can be categorized two kinds of solutions: pipeline and end-to-end detector . However, all these existing methods struggle to detect impersonation audios. Thus,we propose a novel framework that integrate speaker-specific profile into the detection of the fake audios that are imitated by actual human beings. Also we present the first large-scale Chinese impersonation dataset IPAD (ImPersonation Audio Detection) to benefit the community's research on impersonation audio detection. The impersonation dataset will be public.
Submission Number: 4946
Loading