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.

Hand pose estimation from single depth images is an essential topic in computer vision and human computer interaction. Despite recent advancements in this area promoted by convolutional neural networks, accurate hand pose estimation is still a challenging problem. In this paper we propose a novel approach named as Pose guided structured Region Ensemble Network (Pose-REN) to boost the performance of hand pose estimation. Under the guidance of an initially estimated pose, the proposed method extracts regions from the feature maps of convolutional neural network and generates more optimal and representative features for hand pose estimation. The extracted feature regions are then integrated hierarchically according to the topology of hand joints by tree-structured fully connections to regress the refined hand pose. The final hand pose is obtained by an iterative cascaded method. Comprehensive experiments on public hand pose datasets demonstrate that our proposed method outperforms state-of-the-art algorithms.

Long-range corrected hybrids represent an increasingly popular class of functionals for density functional theory (DFT) that have proven to be very successful for a wide range of chemical applications. In this Communication, we examine the performance of these functionals for time-dependent (TD)DFT descriptions of triplet excited states. Our results reveal that the triplet energies are particularly sensitive to the range-separation parameter; this sensitivity can be traced back to triplet instabilities in the ground state coming from the large effective amounts of Hartree-Fock exchange included in these functionals. As such, the use of standard long-range corrected functionals for the description of triplet states at the TDDFT level is not recommended.

Electronic delocalization effects have been proposed to play a key role in photocurrent generation in organic photovoltaic devices. Here, we study the role of charge delocalization on the nature of the charge-transfer (CT) states in the case of model complexes consisting of several pentacene molecules and one fullerene (C60) molecule, which are representative of donor/acceptor heterojunctions. The energies of the CT states are examined by means of time-dependent density functional theory (TD-DFT) using the long-range-corrected functional, ωB97X, with an optimized range-separation parameter, ω. We provide a general description of how the nature of the CT states is impacted by molecular packing (i.e., interfacial donor/acceptor orientations), system size, and intermolecular interactions, features of importance in the understanding of the charge-separation mechanism.