Picture for Wenyuan Xu

Wenyuan Xu

Zhejiang University

SOPHON: Non-Fine-Tunable Learning to Restrain Task Transferability For Pre-trained Models

Add code
Apr 19, 2024
Viaarxiv icon

SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models

Add code
Apr 10, 2024
Figure 1 for SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models
Figure 2 for SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models
Figure 3 for SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models
Figure 4 for SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models
Viaarxiv icon

CamPro: Camera-based Anti-Facial Recognition

Add code
Dec 30, 2023
Viaarxiv icon

TPatch: A Triggered Physical Adversarial Patch

Add code
Dec 30, 2023
Viaarxiv icon

Inaudible Adversarial Perturbation: Manipulating the Recognition of User Speech in Real Time

Add code
Aug 03, 2023
Figure 1 for Inaudible Adversarial Perturbation: Manipulating the Recognition of User Speech in Real Time
Figure 2 for Inaudible Adversarial Perturbation: Manipulating the Recognition of User Speech in Real Time
Figure 3 for Inaudible Adversarial Perturbation: Manipulating the Recognition of User Speech in Real Time
Figure 4 for Inaudible Adversarial Perturbation: Manipulating the Recognition of User Speech in Real Time
Viaarxiv icon

Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound

Add code
Jun 28, 2023
Figure 1 for Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound
Figure 2 for Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound
Figure 3 for Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound
Figure 4 for Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound
Viaarxiv icon

Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion

Add code
Feb 24, 2023
Figure 1 for Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion
Figure 2 for Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion
Figure 3 for Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion
Figure 4 for Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion
Viaarxiv icon

V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization

Add code
Oct 27, 2022
Figure 1 for V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization
Figure 2 for V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization
Figure 3 for V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization
Figure 4 for V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization
Viaarxiv icon

Private Eye: On the Limits of Textual Screen Peeking via Eyeglass Reflections in Video Conferencing

Add code
May 08, 2022
Figure 1 for Private Eye: On the Limits of Textual Screen Peeking via Eyeglass Reflections in Video Conferencing
Figure 2 for Private Eye: On the Limits of Textual Screen Peeking via Eyeglass Reflections in Video Conferencing
Figure 3 for Private Eye: On the Limits of Textual Screen Peeking via Eyeglass Reflections in Video Conferencing
Figure 4 for Private Eye: On the Limits of Textual Screen Peeking via Eyeglass Reflections in Video Conferencing
Viaarxiv icon

Rolling Colors: Adversarial Laser Exploits against Traffic Light Recognition

Apr 06, 2022
Figure 1 for Rolling Colors: Adversarial Laser Exploits against Traffic Light Recognition
Figure 2 for Rolling Colors: Adversarial Laser Exploits against Traffic Light Recognition
Figure 3 for Rolling Colors: Adversarial Laser Exploits against Traffic Light Recognition
Figure 4 for Rolling Colors: Adversarial Laser Exploits against Traffic Light Recognition
Viaarxiv icon