Hand gestures are a natural way for human-robot interaction.Vision based dynamic hand gesture recognition has become a hot research topic due to its various applications.This paper presents a novel deep learning netwo...Hand gestures are a natural way for human-robot interaction.Vision based dynamic hand gesture recognition has become a hot research topic due to its various applications.This paper presents a novel deep learning network for hand gesture recognition.The network integrates several well-proved modules together to learn both short-term and long-term features from video inputs and meanwhile avoid intensive computation.To learn short-term features,each video input is segmented into a fixed number of frame groups.A frame is randomly selected from each group and represented as an RGB image as well as an optical flow snapshot.These two entities are fused and fed into a convolutional neural network(Conv Net)for feature extraction.The Conv Nets for all groups share parameters.To learn longterm features,outputs from all Conv Nets are fed into a long short-term memory(LSTM)network,by which a final classification result is predicted.The new model has been tested with two popular hand gesture datasets,namely the Jester dataset and Nvidia dataset.Comparing with other models,our model produced very competitive results.The robustness of the new model has also been proved with an augmented dataset with enhanced diversity of hand gestures.展开更多
The coaxial surface wave linear plasma with preeminent axial uniformity is developed with the 2.45 GHz microwave generator.By optical emission spectroscopy,parameters of the argon linear plasma with a length over 600 ...The coaxial surface wave linear plasma with preeminent axial uniformity is developed with the 2.45 GHz microwave generator.By optical emission spectroscopy,parameters of the argon linear plasma with a length over 600 mm are diagnosed under gas pressure of 30 and 50 Pa and different microwave powers.The spectral lines of argon and Hβ(486.1 nm)atoms in excited state are observed for estimating electron excitation temperature and electron density.Spectrum bands in305–310 nm of diatomic OH(Σ-Π+A X22 i)radicals are used to determine the molecule rotational temperature.Finally,the axial uniformity of electron density and electron excitation temperature are analyzed emphatically under various conditions.The results prove the distinct optimization of compensation from dual powers input,which can narrow the uniform coefficient of electron density and electron excitation temperature by around 40%and 22%respectively.With the microwave power increasing,the axial uniformity of both electron density and electron excitation temperature performs better.Nevertheless,the fluctuation of electron density along the axial direction appeared with higher gas pressure.The axial uniformity of coaxial surface wave linear plasma could be controlled by pressure and power for a better utilization in material processing.展开更多
Two-dimensional （2D） layered materials, transition-metal dichalcogenides, and black phosphorus have attracted considerable interest from the viewpoints of fundamental physics and device applications. The establishme...Two-dimensional （2D） layered materials, transition-metal dichalcogenides, and black phosphorus have attracted considerable interest from the viewpoints of fundamental physics and device applications. The establishment of new functionalities in anisotropic layered 2D materials is a challenging but rewarding frontier, owing to the remarkable optical properties of these materials and their prospects for new devices. Herein, we report the anisotropic and thickness- dependent optical properties of a 2D layered monochalcogenide of germanium sulfide （GeS）. Three Raman-scattering peaks corresponding to the B3g,, A1g, and A2g modes with a strong polarization dependence are demonstrated in the GeS flakes, which validates polarized Raman spectroscopy as an effective method for identifying the crystal orientation of anisotropic layered GeS. Photoluminescence （PL） is observed with a peak at -1.66 eV that originates from the direct optical transition in GeS at room temperature. The polarization-dependent characteristics of the PL, which are revealed for the first time, along with the demonstration of anisotropic absorption, indicate an obvious anisotropic optical transition near the band edge of GeS, which is supported by density functional theory calculations. The significantly thickness-dependent PL is observed and discussed. This anisotropic layered GeS presents opportunities for the discovery of new physical phenomena and will find applications that exploit its anisotropic properties, such as polarization-sensitive photodetectors.展开更多
1 Introduction Visual tracking has long been considered as a fundamental mission in computer vision,and it has been widely used for many applications.Siamese trackers[1]utilize Siamese network to do template matching ...1 Introduction Visual tracking has long been considered as a fundamental mission in computer vision,and it has been widely used for many applications.Siamese trackers[1]utilize Siamese network to do template matching and achieve state-of-the-art performance recently.However,the Siamese trackers fail to distinguish the target from the distractors with the same category and perform unstably in complicated environment.We find there are three reasons which affect the robustness and stability of the tracker.展开更多
1 Introduction Referring image segmentation aims to identify and segment the instance by language description.It is a challenge task due to the combination of both language and visual domains[1].The current mainstream...1 Introduction Referring image segmentation aims to identify and segment the instance by language description.It is a challenge task due to the combination of both language and visual domains[1].The current mainstream methods mainly focus on English,and can achieve an appropriated performance.However,it cannot achieve the similar performance while they are directly applied to semantic-oriented languages like Chinese due to the different language characteristics.展开更多
文摘Hand gestures are a natural way for human-robot interaction.Vision based dynamic hand gesture recognition has become a hot research topic due to its various applications.This paper presents a novel deep learning network for hand gesture recognition.The network integrates several well-proved modules together to learn both short-term and long-term features from video inputs and meanwhile avoid intensive computation.To learn short-term features,each video input is segmented into a fixed number of frame groups.A frame is randomly selected from each group and represented as an RGB image as well as an optical flow snapshot.These two entities are fused and fed into a convolutional neural network(Conv Net)for feature extraction.The Conv Nets for all groups share parameters.To learn longterm features,outputs from all Conv Nets are fed into a long short-term memory(LSTM)network,by which a final classification result is predicted.The new model has been tested with two popular hand gesture datasets,namely the Jester dataset and Nvidia dataset.Comparing with other models,our model produced very competitive results.The robustness of the new model has also been proved with an augmented dataset with enhanced diversity of hand gestures.
基金supported by National Natural Science Foundation of China(Nos.11575252 and 11775270)Institute of Energy of Hefei Comprehensive National Science Center,People’s Republic of China(Nos.19KZS206,21KZS201)。
文摘The coaxial surface wave linear plasma with preeminent axial uniformity is developed with the 2.45 GHz microwave generator.By optical emission spectroscopy,parameters of the argon linear plasma with a length over 600 mm are diagnosed under gas pressure of 30 and 50 Pa and different microwave powers.The spectral lines of argon and Hβ(486.1 nm)atoms in excited state are observed for estimating electron excitation temperature and electron density.Spectrum bands in305–310 nm of diatomic OH(Σ-Π+A X22 i)radicals are used to determine the molecule rotational temperature.Finally,the axial uniformity of electron density and electron excitation temperature are analyzed emphatically under various conditions.The results prove the distinct optimization of compensation from dual powers input,which can narrow the uniform coefficient of electron density and electron excitation temperature by around 40%and 22%respectively.With the microwave power increasing,the axial uniformity of both electron density and electron excitation temperature performs better.Nevertheless,the fluctuation of electron density along the axial direction appeared with higher gas pressure.The axial uniformity of coaxial surface wave linear plasma could be controlled by pressure and power for a better utilization in material processing.
文摘Two-dimensional （2D） layered materials, transition-metal dichalcogenides, and black phosphorus have attracted considerable interest from the viewpoints of fundamental physics and device applications. The establishment of new functionalities in anisotropic layered 2D materials is a challenging but rewarding frontier, owing to the remarkable optical properties of these materials and their prospects for new devices. Herein, we report the anisotropic and thickness- dependent optical properties of a 2D layered monochalcogenide of germanium sulfide （GeS）. Three Raman-scattering peaks corresponding to the B3g,, A1g, and A2g modes with a strong polarization dependence are demonstrated in the GeS flakes, which validates polarized Raman spectroscopy as an effective method for identifying the crystal orientation of anisotropic layered GeS. Photoluminescence （PL） is observed with a peak at -1.66 eV that originates from the direct optical transition in GeS at room temperature. The polarization-dependent characteristics of the PL, which are revealed for the first time, along with the demonstration of anisotropic absorption, indicate an obvious anisotropic optical transition near the band edge of GeS, which is supported by density functional theory calculations. The significantly thickness-dependent PL is observed and discussed. This anisotropic layered GeS presents opportunities for the discovery of new physical phenomena and will find applications that exploit its anisotropic properties, such as polarization-sensitive photodetectors.
基金the National Natural Science Foundation of China(Grant No.62076246).
文摘1 Introduction Visual tracking has long been considered as a fundamental mission in computer vision,and it has been widely used for many applications.Siamese trackers[1]utilize Siamese network to do template matching and achieve state-of-the-art performance recently.However,the Siamese trackers fail to distinguish the target from the distractors with the same category and perform unstably in complicated environment.We find there are three reasons which affect the robustness and stability of the tracker.
基金supported in part by National Natural Science Foundation of China (Grant No.62076246).
文摘1 Introduction Referring image segmentation aims to identify and segment the instance by language description.It is a challenge task due to the combination of both language and visual domains[1].The current mainstream methods mainly focus on English,and can achieve an appropriated performance.However,it cannot achieve the similar performance while they are directly applied to semantic-oriented languages like Chinese due to the different language characteristics.
