Scholarlink

<p>Recently, rise in voice spoofing attacks have posed a significant threat to the trustworthiness and reliability of Automatic Speaker Verification (ASV) systems. Such attacks can lead to the spread of misinformation. In recent years, numerous methods have been developed by various researchers to safeguard ASV systems from spoofing attacks. With the substantial progress in the field, there remains a critical need to organize and categorize research such that qualitative and quantitative comparisons of state-of-the-art countermeasures can be performed. The existing literature reveals that most of the survey papers have reviewed the existing methods by considering only front-end feature extraction techniques and back-end classification methods. There have been very less efforts to conduct an unbiased comparison of published countermeasures’ performance and their generalizability across various types of attacks. Also, no comprehensive survey has covered all types of spoofing attacks such as synthetic, replay, and deep fake attacks. This paper aims to bridge this gap by presenting an all-encompassing survey that includes all types of voice-spoofing attacks, feature extraction techniques, classification methods, datasets and evaluation metrics. Firstly, the paper presents the taxonomy of attacks along brief discussion of various other parameters chosen for making the comparison such as front-end feature extraction methods, backend classification techniques, datasets, and evaluation metrics. Secondly, it presents an analysis of various existing countermeasures by dividing them into two categories such as countermeasures that have taken only one of the three types of spoofing attacks into consideration, and countermeasures that have taken more than one of the three types of spoofing attacks into consideration. Also, while doing the review, the paper discusses the contributions and limitations faced by the current spoofing countermeasures. Finally, the paper discusses various challenges that still exist and provides directions for future research in this important area.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Review and analysis of audio spoof countermeasures: an attack perspective

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Nidhi Chakravarty,
  • Mohit Dua

摘要

Recently, rise in voice spoofing attacks have posed a significant threat to the trustworthiness and reliability of Automatic Speaker Verification (ASV) systems. Such attacks can lead to the spread of misinformation. In recent years, numerous methods have been developed by various researchers to safeguard ASV systems from spoofing attacks. With the substantial progress in the field, there remains a critical need to organize and categorize research such that qualitative and quantitative comparisons of state-of-the-art countermeasures can be performed. The existing literature reveals that most of the survey papers have reviewed the existing methods by considering only front-end feature extraction techniques and back-end classification methods. There have been very less efforts to conduct an unbiased comparison of published countermeasures’ performance and their generalizability across various types of attacks. Also, no comprehensive survey has covered all types of spoofing attacks such as synthetic, replay, and deep fake attacks. This paper aims to bridge this gap by presenting an all-encompassing survey that includes all types of voice-spoofing attacks, feature extraction techniques, classification methods, datasets and evaluation metrics. Firstly, the paper presents the taxonomy of attacks along brief discussion of various other parameters chosen for making the comparison such as front-end feature extraction methods, backend classification techniques, datasets, and evaluation metrics. Secondly, it presents an analysis of various existing countermeasures by dividing them into two categories such as countermeasures that have taken only one of the three types of spoofing attacks into consideration, and countermeasures that have taken more than one of the three types of spoofing attacks into consideration. Also, while doing the review, the paper discusses the contributions and limitations faced by the current spoofing countermeasures. Finally, the paper discusses various challenges that still exist and provides directions for future research in this important area.