On The Generation and Removal of Speaker Adversarial Perturbation For Voice-Privacy Protection

Keyphrases

• Privacy Protection 100%
• Adversarial Perturbation 100%
• Voice Privacy 100%
• Anonymized 66%
• Neural Network 33%
• Removal Methods 33%
• Security Analysis 33%
• Insecurity 33%
• Joint Training 33%
• Adversarial Attack 33%
• Adversarial Network 33%
• Forensic Analysis 33%
• Deanonymize 33%
• LibriSpeech 33%
• Perturbation Signals 33%
• Additive Perturbation 33%
• Reversibility Property 33%
• Criminal Identity 33%

Computer Science

• Privacy Protection 100%
• Use Case 33%
• Adversarial Machine Learning 33%
• Neural Network 33%
• Experimental Result 33%
• Removal Process 33%