arXiv subjects

Xinghao Chen

Publications and source records attributed to Xinghao Chen.

114 records · Page 7Linked to original sources

arXiv2017

Motion Feature Augmented Recurrent Neural Network for Skeleton-based Dynamic Hand Gesture Recognition

Dynamic hand gesture recognition has attracted increasing interests because of its importance for human computer interaction. In this paper, we propose a new motion feature augmented recurrent neural network for skeleton-based dynamic hand gesture recognition. Finger motion features are extracted to describe finger movements and global motion features are utilized to represent the global movement of hand skeleton. These motion features are then fed into a bidirectional recurrent neural network (RNN) along with the skeleton sequence, which can augment the motion features for RNN and improve the classification performance. Experiments demonstrate that our proposed method is effective and outperforms start-of-the-art methods.

cs.CV↗

arXiv2017

Towards Good Practices for Deep 3D Hand Pose Estimation

3D hand pose estimation from single depth image is an important and challenging problem for human-computer interaction. Recently deep convolutional networks (ConvNet) with sophisticated design have been employed to address it, but the improvement over traditional random forest based methods is not so apparent. To exploit the good practice and promote the performance for hand pose estimation, we propose a tree-structured Region Ensemble Network (REN) for directly 3D coordinate regression. It first partitions the last convolution outputs of ConvNet into several grid regions. The results from separate fully-connected (FC) regressors on each regions are then integrated by another FC layer to perform the estimation. By exploitation of several training strategies including data augmentation and smooth $L_1$ loss, proposed REN can significantly improve the performance of ConvNet to localize hand joints. The experimental results demonstrate that our approach achieves the best performance among state-of-the-art algorithms on three public hand pose datasets. We also experiment our methods on fingertip detection and human pose datasets and obtain state-of-the-art accuracy.

cs.CV↗

arXiv2017

Region Ensemble Network: Improving Convolutional Network for Hand Pose Estimation

Hand pose estimation from monocular depth images is an important and challenging problem for human-computer interaction. Recently deep convolutional networks (ConvNet) with sophisticated design have been employed to address it, but the improvement over traditional methods is not so apparent. To promote the performance of directly 3D coordinate regression, we propose a tree-structured Region Ensemble Network (REN), which partitions the convolution outputs into regions and integrates the results from multiple regressors on each regions. Compared with multi-model ensemble, our model is completely end-to-end training. The experimental results demonstrate that our approach achieves the best performance among state-of-the-arts on two public datasets.

cs.CV↗

arXiv2017

Rotational splitting and asteroseismic modelling of the delta Scuti star EE Camelopardalis

According to the rotational splitting law of g modes, the frequency spectra of EE Cam can be disentangled only with oscillation modes of $\ell$ = 0, 1, and 2. Fifteen sets of rotational splits are found, containing five sets of $\ell=1$ multiplets and ten sets of $\ell=2$ multiplets. The rotational period of EE Cam is deduced to be $P_{\rm rot}$ = $1.84_{-0.05}^{+0.07}$ days. When we do model fittings, we use two nonradial oscillation modes ($f_{11}$ and $f_{32}$) and the fundamental radial mode $f_{1}$. The fitting results show that $χ^{2}$ of the best-fitting model is much smaller than those of other theoretical models. The physical parameters of the best-fitting model are $M$ = 2.04 $M_{\odot}$, $Z=0.028$, $T_{\rm eff}$ = 6433 K, $\log L/L_{\odot}$ = 1.416, $R$ = 4.12 $R_{\odot}$, $\log g$ = 3.518, and $χ^{2}$ = 0.00035. Furthermore, we find $f_{11}$ and $f_{32}$ are mixed mode, which mainly characterize the features of the helium core. The fundamental radial mode $f_{1}$ mainly restrict the features of the stellar envelope. Finally, the acoustic radius $τ_{0}$ and the period separation $Π_{0}$ are determined to be 5.80 hr and 463.7 s respectively, and the size of the helium core of EE Cam is estimated to be $M_{\rm He}$ = 0.181 $M_{\odot}$ and $R_{\rm He}$ = 0.0796 $R_{\odot}$.

astro-ph.SR↗

arXiv2016

Two-stream convolutional neural network for accurate RGB-D fingertip detection using depth and edge information

Accurate detection of fingertips in depth image is critical for human-computer interaction. In this paper, we present a novel two-stream convolutional neural network (CNN) for RGB-D fingertip detection. Firstly edge image is extracted from raw depth image using random forest. Then the edge information is combined with depth information in our CNN structure. We study several fusion approaches and suggest a slow fusion strategy as a promising way of fingertip detection. As shown in our experiments, our real-time algorithm outperforms state-of-the-art fingertip detection methods on the public dataset HandNet with an average 3D error of 9.9mm, and shows comparable accuracy of fingertip estimation on NYU hand dataset.

cs.CV↗

arXiv2016

Exploring the helium core of the delta Scuti star CoRoT 102749568 with asteroseismology

Based on regularities in rotational splittings, we seek possible multiplets for the observed frequencies of CoRoT 102749568. Twenty-one sets of multiplets are identified, including four sets of multiplets with $l=1$, nine sets of multiplets with $l=2$, and eight sets of multiplets with $l=3$. In particular, there are three complete triplets ($f_{10}$, $f_{12}$, $f_{14}$), ($f_{31}$, $f_{34}$, $f_{35}$), and ($f_{41}$, $f_{43}$, $f_{44}$). The rotational period of CoRoT 102749568 is estimated to be $1.34^{+0.04}_{-0.05}$ days. When doing model fittings, three $l=1$ modes ($f_{12}$, $f_{34}$, and $f_{43}$) and the radial first overtone $f_{13}$ are used. Our results shows that the three nonradial modes ($f_{12}$, $f_{34}$, and $f_{43}$) are mixed modes, which mainly provide constraints on the helium core. The radial first overtone $f_{13}$ mainly provides constraint on the stellar envelope. Hence the size of the helium core of CoRoT 102749568 is determined to be $M_{\rm He}$ = 0.148 $\pm$ 0.003 $M_{\odot}$ and $R_{\rm He}$ = 0.0581 $\pm$ 0.0007 $R_{\odot}$. The fundamental parameters of CoRoT 102749568 are determined to be $M$ = 1.54 $\pm$ 0.03 $M_{\odot}$, $Z=$ 0.006, $f_{\rm ov}$ = 0.004 $\pm$ 0.002, $\log g$ = 3.696 $\pm$ 0.003, $T_{\rm eff}$ = 6886 $\pm$ 70 K, $R$ = 2.916 $\pm$ 0.039 $R_{\odot}$, and $L$ = 17.12 $\pm$ 1.13 $L_{\odot}$.

astro-ph.SR↗