SYSTEMS AND DEVICES TO INFER HAND GESTURES AND OTHER HAND INTERACTIONS

Smart overview of the Invention

Wearable devices designed to interpret hand gestures and interactions are discussed, focusing on two main forms: a ring-shaped device and a wrist-worn device. These devices employ sensors that capture acoustic signals and depth information, processed by an onboard processor to detect hand movements. The invention aims to enhance gesture recognition by addressing common issues like sensor-shift and privacy concerns associated with camera-based techniques.

Technical Background

Hand gesture recognition has been a significant area of research, with various methods proposed, including inertial measurement units and electrical techniques. However, camera-based methods, despite their accuracy, face challenges like high processing demands, occlusion, and privacy issues. The disclosed wearable devices seek to overcome these limitations by using acoustic and depth sensors, providing a more efficient and privacy-conscious solution.

Device Functionality

The devices incorporate sensors to acquire depth images and bioacoustic signals, which are processed to interpret hand gestures. A depth sensor captures low-resolution images to create a 3D hand model, while an acoustic device records vibrations through the hand. Machine learning algorithms then infer gestures or hand states from this combined data. The wrist-worn device, in particular, uses a Time-of-Flight (TOF) sensor and a vibration sensor to enhance gesture-based controls for smart devices.

Design and Calibration

The wrist-worn device features a TOF sensor and a Voice Pickup Unit (VPU) for capturing depth and acoustic data, respectively. It addresses the sensor-shift issue by implementing a calibration algorithm that maintains consistent sensor positioning relative to the hand. This ensures accurate gesture tracking and enhances user experience by offering a reliable and privacy-focused interface for gesture-based applications.

Applications and Advantages

These wearable devices are designed to facilitate intuitive gesture-based control for smart devices and cameras. By combining depth and bioacoustic information, the devices can accurately interpret dynamic gestures and microgestures, even in mid-air or when interacting with objects. The compact, wrist-worn design makes it practical for everyday use, while the limited resolution and field of view of the sensors help mitigate privacy concerns.