Hu proposed a method that recognizes finger movements by detecting wrist muscles for human-computer interaction (HCI). Considering human habits and aesthetics, the sensor is placed on the back of the wrist. We first designed a polyvinylidene fluoride (PVDF) piezoelectric thin-film sensor unit with a planar elastic substrate. By studying the effects of the hardness and thickness of the substrate, we designed the sensor unit to have a Shore-A hardness of 23 and a thickness of 2 mm. Then we constructed a 4 × 2 sensor matrix with a size of 25 × 15 mm 2 . To build a finger movement dataset, we collected wrist dorsal movement signals when the fingers moved using the sensor matrix. Then we used a four-layer back-propagation (BP) neural network to recognize the finger movements. We experimentally demonstrated that even on the dorsal wrist side, finger movements could be recognized. The recognition rate of the general model using mixed personal data was 79%. In comparison, the recognition rate of the individual model using personal data was 94%.

Depth and ego-motion estimations are essential for the localization and navigation of autonomous robots and autonomous driving. Recent studies make it possible to learn the per-pixel depth and ego-motion from the unlabeled monocular video. A novel unsupervised training framework is proposed with 3D hierarchical refinement and augmentation using explicit 3D geometry. In this framework, the depth and pose estimations are hierarchically and mutually coupled to refine the estimated pose layer by layer. The intermediate view image is proposed and synthesized by warping the pixels in an image with the estimated depth and coarse pose. Then, the residual pose transformation can be estimated from the new view image and the image of the adjacent frame to refine the coarse pose. The iterative refinement is implemented in a differentiable manner in this paper, making the whole framework optimized uniformly. Meanwhile, a new image augmentation method is proposed for the pose estimation by synthesizing a new view image, which creatively augments the pose in 3D space but gets a new augmented 2D image. The experiments on KITTI demonstrate that our depth estimation achieves state-of-the-art performance and even surpasses recent approaches that utilize other auxiliary tasks. Our visual odometry outperforms all recent unsupervised monocular learning-based methods and achieves competitive performance to the geometrybased method, ORB-SLAM2 with back-end optimization.