论文标题
Faceori：使用耳机上的超声范围跟踪头部位置和方向
FaceOri: Tracking Head Position and Orientation Using Ultrasonic Ranging on Earphones
论文作者
论文摘要
面部方向通常表明用户的预期交互目标。在本文中，我们提出了Faceori，这是一种基于使用耳机的声学范围的新型面部跟踪技术。 Faceori可以利用商品设备上的扬声器发出超声波呼叫，该呼声是由用户耳机上的麦克风集接收的，然后处理以计算从每个麦克风到设备的距离。这些测量用于得出用户的面向方向和相对于设备的距离。我们进行了基础真理比较和用户研究，以评估Faceori的表现。结果表明，系统可以确定用户是否以1.5米范围内的93.5％精度对设备的构成。此外，Faceori可以通过距离的中值绝对误差为10.9 mm，偏航为3.7度，在螺距中的5.8度，可以连续跟踪用户的头部方向。 Faceori可以轻松地控制设备，并产生更智能的上下文感知互动。
Face orientation can often indicate users' intended interaction target. In this paper, we propose FaceOri, a novel face tracking technique based on acoustic ranging using earphones. FaceOri can leverage the speaker on a commodity device to emit an ultrasonic chirp, which is picked up by the set of microphones on the user's earphone, and then processed to calculate the distance from each microphone to the device. These measurements are used to derive the user's face orientation and distance with respect to the device. We conduct a ground truth comparison and user study to evaluate FaceOri's performance. The results show that the system can determine whether the user orients to the device at a 93.5% accuracy within a 1.5 meters range. Furthermore, FaceOri can continuously track the user's head orientation with a median absolute error of 10.9 mm in the distance, 3.7 degrees in yaw, and 5.8 degrees in pitch. FaceOri can allow for convenient hands-free control of devices and produce more intelligent context-aware interaction.
