Megatron: Evasive Clean-Label Backdoor Attacks against Vision Transformer

Xueluan Gong, Bowei Tian, Meng Xue^(Corresponding), Shuaike Li, Yanjiao Chen, Qian Wang

• In this work, we propose MEGATRON, an evasive clean-label backdoor attack against vision transformers, where the attacker injects the backdoor without manipulating the data-labeling process.

Home-based Dry Eye Assessment via Blink Kinematics Using mmWave and Clinical Knowledge Distillation

Meng Xue, Wentao Xie, Zuohuizi Yi, Zhilong Zhang, Shumao Wu, Yinan Zhu, Qian Zhang, Changzheng Chen

• In this work, we present Blinic, a contactless system leveraging commercial millimeter-wave (mmWave) radar to predict precise TBUT values and assess DED severity grades at home.

DD-LIVM: Pioneering Cross-Domain Photovoltaic Defect Detection Using Large Infrared-Visible Model

Yinan Zhu, Meng Xue, Haiyan Hu, Cong Zhang, Xiaoyi Fan, Qian Zhang

• In this work, we present the first large infrared-visible model DD-LIVM to enable cross-domain defect detection.

MobHAR: Source-free Knowledge Transfer for Human Activity Recognition on Mobile Devices

Meng Xue, Yinan Zhu, Wentao Xie, Zhixian Wang, Yanjiao Chen, Kui Jiang, and Qian Zhang

• Code
• In this work, we present a novel source-free domain adaptation framework for HAR that effectively handles substantial domain discrepancies across different datasets. Our approach introduces two key innovations: (1) a Discriminative Information Gramian (DIG) method that quantifies the relationship between target-domain samples and the source domain without requiring access to source data, and (2) an unsupervised domain generalization technique that ensures consistent feature extraction across augmented data samples, enhancing the model’s effectiveness in the target domain.

SDE: Early Screening for Dry Eye Disease with Wireless Signals

Meng Xue, Yuyang Zeng, Shengkang Gu, Qian Zhang, Bowei Tian, and Changzheng Chen

• Demo

• In this paper, we propose SDE, a contactless, convenient, and ubiquitous DED screening system based on RF signals. By constructing frame chirps variance, we extract fine-grained spontaneous blinking action from RF signals.

ARTEMIS: Defending against Backdoor Attacks via Distribution Shift

Meng Xue, Zhixian Wang, Qian Zhang, Xueluan Gong, Zhihang Liu, and Yanjiao Chen

• Code
• In this work, we propose a novel backdoor defense approach called ARTEMIS, which utilizes distribution shifts to eliminate the discrepancy between poisoned and benign samples in the feature space. Additionally, ARTEMIS learns from domain-invariant features after the shift.

Echo: Reverberation-based Fast Black-Box Adversarial Attacks on Intelligent Audio Systems

Meng Xue, Peng Kuang, Xueluan Gong, Qian ZHANG, and Yanjiao Chen

• Projects
• In this paper, we propose a physical adversarial attack that exploits reverberation, a natural indoor acoustic effect, to realize imperceptible, fast, and targeted black-box attacks. Moreover, we generate reverberation-alike perturbations that blend naturally with the original voice sample.

Wet-Ra: Monitoring Diapers Wetness with Wireless Signals

Meng Xue, Yanjiao Chen, Xueluan Gong, Jian Zhang, and Chunkai Fan

• Presentaion
• Diaper wetness monitoring is essential in various situations (e.g., babies and patients) to guarantee hygiene and avoid embarrassment. In this paper, we introduce Wet-Ra, a contactless, ubiquitous, and user-friendly diaper wetness monitoring system based on RF signals.

ChestLive: Fortifying Voice-based Authentication with Chest Motion Biometric on Smart Devices

Yanjiao Chen, Meng Xue, Jian Zhang, Qianyun Guan, Zhiyuan Wang, Qian Zhang, and Wei Wang

• In this paper, we propose to leverage the distinctive chest motions during speaking to establish a secure multi-factor authentication system, named ChestLive. We use acoustic sensing to monitor chest motions with a built-in speaker and microphone on smartphones.