APFT: Adaptive Phoneme Filter Template to Generate Anti-Compression Speech Adversarial Example in Real-Time

Automatic Speech Recognition (ASR) systems are widely used for speech censoring. Speech Adversarial Example (AE) offers a novel approach to protect speech privacy by forcing ASR to mistranscribe. However, existing speech AE faces two challenges in real-time voice communication scenarios, such as IP telephone, voice chat, or video conference, it cannot be generated in real-time, and its defensive capability is significantly reduced after the essential audio compression for network transmission. In this paper, we propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Adaptive Phoneme Filter Template (APFT)</i> to address these issues. The key features of APFT include: 1) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Phoneme-level Templates</i> for universal AE generation in real-time, 2) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Filter</i>, which eliminates redundant signals to improve compression robustness. 3) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Adaptive Band Filtering</i>, which limits the attack area from the frequency band without affecting the attack effectiveness and improves speech quality. The comprehensive experimental results show that APFT has four advantages: 1) Real-time Generation, with AE generation time below 1.1ms for 1s speech; 2) Compression Robustness, achieving a WER of 0.64 under AAC and Opus codecs; 3) Transferability, with an average WER of 0.72 across datasets and ASR systems; 4) Stealthiness, achieving a MOS of 4.07 for high-quality speech. In addition, the experiment on Telegram voice calls further proves the practical applicability of APFT. The demo of APFT can be obtained in https://yihuan-qaq.github.io/APFT.github.io/.