Abstract In this paper, we design and fabricate a Schottky-metal-insulator-semiconductor (MIS) cascode anode GaN lateral field-effect diode (CA-LFED) to achieve ultralow reverse leakage current ( I LEAK ). The device based on AlGaN/GaN high-electron-mobility-transistor (HEMT) includes a normally-off MIS-controlled channel that is cascoded with a high barrier height Schottky contact. At reverse bias, the high electric-field is effectively prevented by the normally-off MIS-controlled channel edge. Together with the high barrier height Schottky contact, this feature significantly suppresses the I LEAK . Supported by the device fabrication, the CA-LFED with high breakdown voltage (BV) > 600 V shows an ultralow I LEAK of 3.6 × 10 −9 A/mm as well as a low forward voltage drop ( V F ) of 2.2 V. The performance suggests that the CA-LFED can be a promising candidate for ultralow I LEAK and better V F - I LEAK trade-off GaN power diode applications.

Most existing sign language recognition works are based on videos, which have the characteristics of weak robustness and high computational complexity. In this paper, we propose a new continuous sign language recognition (CSLR) framework to translate sign language represented by multi-part skeleton data into text. In the framework, the Multi-Stream Spatial-Temporal Graph Convolutional Networks (Multi-Stream ST-GCNs) is designed to model dynamic multi-part skeletons with graph structure. Meanwhile, we utilize adaptive fusion to weighted fuse the features of different parts, which can highlight the critical parts and weaken the auxiliary parts. The Attention-based Encoder-Decoder Network is applied to learn the mapping between the fused feature sequence and target sentence. In addition, we build the first skeleton-based CSLR dataset using Leap Motion, named LMSLR. Experimental results demonstrate that our approach achieves advanced performance in word error rate (WER).