Publications

2024

1. 

   SonicID: User Identification on Smart Glasses with Acoustic Sensing

   K Li, Devansh Agarwal, R Zhang, V Gunda, T Mo, S Mahmud, B Chen, F Guimbretière, and C Zhang

   Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 2024

   SonicID enables user identification on smart glasses using acoustic sensing techniques. Our system can distinguish between different users based on their unique acoustic signatures captured through the glasses.

2. 

   MunchSonic: Tracking Fine-grained Dietary Actions through Active Acoustic Sensing on Eyeglasses

   S Mahmud, Devansh Agarwal, A Ajit, Q Liang, T Viranda, F Guimbretière, and C Zhang

   In Proceedings of the 2024 ACM International Symposium on Wearable Computers, 2024

   MunchSonic uses active acoustic sensing on eyeglasses to track fine-grained dietary actions. The system can detect and classify different eating behaviors with high accuracy, enabling detailed dietary monitoring.

3. 

   EchoGuide: Active Acoustic Guidance for LLM-Based Eating Event Analysis from Egocentric Videos

   V Parikh, S Mahmud, Devansh Agarwal, K Li, F Guimbretière, and C Zhang

   In Proceedings of the 2024 ACM International Symposium on Wearable Computers, 2024

   Best Paper Honorable Mention

   Best Paper Honorable Mention

   EchoGuide combines active acoustic sensing with large language models to provide enhanced eating event analysis from egocentric videos. The system improves accuracy of dietary monitoring through multimodal sensing.

4. 

   Echowrist: Continuous Hand Pose Tracking and Hand-Object Interaction Recognition Using Low-Power Active Acoustic Sensing on a Wristband

   CJ Lee, R Zhang, Devansh Agarwal, TC Yu, V Gunda, O Lopez, J Kim, S Yin, S Nie, F Guimbretière, and C Zhang

   In Proceedings of the CHI Conference on Human Factors in Computing Systems, 2024

   We present Echowrist, a novel wristband device that uses low-power active acoustic sensing to continuously track hand poses and recognize hand-object interactions. Our system achieves high accuracy while maintaining low power consumption.

5. 

   Ring-a-Pose: A Ring for Continuous Hand Pose Tracking

   TC Yu, G Hu, R Zhang, H Lim, S Mahmud, CJ Lee, K Li, Devansh Agarwal, S Nie, F Guimbretière, and C Zhang

   Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 2024

   Ring-a-Pose presents a novel ring-based device for continuous hand pose tracking using acoustic sensing. The system provides accurate hand pose estimation in real-time with minimal power consumption.

6. 

   ActSonic: Everyday Activity Recognition on Smart Glasses using Active Acoustic Sensing

   S Mahmud, V Parikh, Q Liang, K Li, R Zhang, A Ajit, V Gunda, Devansh Agarwal, F Guimbretière, and C Zhang

   Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 2024

   ActSonic presents a comprehensive system for everyday activity recognition on smart glasses using active acoustic sensing. Our approach can recognize a wide range of daily activities with high accuracy.

2023

1. 

   HPSpeech: Silent Speech Interface for Commodity Headphones

   R Zhang, H Chen, Devansh Agarwal, R Jin, K Li, F Guimbretière, and C Zhang

   In Proceedings of the 2023 ACM International Symposium on Wearable Computers, 2023

   Best Paper Honorable Mention

   Best Paper Honorable Mention

   HPSpeech introduces a silent speech interface that works with commodity headphones, enabling speech recognition without vocalizing. Our approach uses acoustic sensing to detect subtle mouth movements and translate them into text.

2021

1. 

   On Adversarial Robustness: A Neural Architecture Search Perspective

   Chaitanya Devaguptapu, Devansh Agarwal, Gaurav Mittal, and Vineeth N Balasubramanian

   IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021

   This paper explores adversarial robustness from a neural architecture search perspective, providing insights into how different architectural choices affect model robustness against adversarial attacks.