Specsync: Enhancing Privacy-Preserving Hand Tracking: An Indirect Sensing Approach Exploration for Wearables

In recent years, the focus on wrist-worn devices for hand gesture and hand pose prediction and estimation has increased in the realm of research. Despite their remarkable accuracy, these devices have raised various concerns, particularly regarding user privacy. My thesis aims to address these concerns by exploring a system that utilizes a privacy-preserving approach to achieve hand tracking. 

To mitigate privacy risks, the proposed system employs Electrical Impedance Tomography (EIT) as its primary sensing technique. EIT is a non-vision-based sensing method that acquires valuable data while ensuring data minimization, thus avoiding unnecessary information capture. This approach allows the device to function independently as a reliable hand pose estimation device.

With the incorporation of EIT, this research aims to enhance both the privacy and usability of wrist-worn devices for hand tracking applications. The proposed system holds large potential for addressing privacy concerns related to existing methods by employing the principles of data minimization through sensing technologies. The results contribute to the development of more secure and privacy preserving technologies for hand pose estimation.