arXiv subjects

Limin Wang

Publications and source records attributed to Limin Wang.

At least 271 records · Page 15Linked to original sources

arXiv2016

Transferring Object-Scene Convolutional Neural Networks for Event Recognition in Still Images

Event recognition in still images is an intriguing problem and has potential for real applications. This paper addresses the problem of event recognition by proposing a convolutional neural network that exploits knowledge of objects and scenes for event classification (OS2E-CNN). Intuitively, it stands to reason that there exists a correlation among the concepts of objects, scenes, and events. We empirically demonstrate that the recognition of objects and scenes substantially contributes to the recognition of events. Meanwhile, we propose an iterative selection method to identify a subset of object and scene classes, which help to more efficiently and effectively transfer their deep representations to event recognition. Specifically, we develop three types of transferring techniques: (1) initialization-based transferring, (2) knowledge-based transferring, and (3) data-based transferring. These newly designed transferring techniques exploit multi-task learning frameworks to incorporate extra knowledge from other networks and additional datasets into the training procedure of event CNNs. These multi-task learning frameworks turn out to be effective in reducing the effect of over-fitting and improving the generalization ability of the learned CNNs. With OS2E-CNN, we design a multi-ratio and multi-scale cropping strategy, and propose an end-to-end event recognition pipeline. We perform experiments on three event recognition benchmarks: the ChaLearn Cultural Event Recognition dataset, the Web Image Dataset for Event Recognition (WIDER), and the UIUC Sports Event dataset. The experimental results show that our proposed algorithm successfully adapts object and scene representations towards the event dataset and that it achieves the current state-of-the-art performance on these challenging datasets.

cs.CV↗

arXiv2016

CUHK & ETHZ & SIAT Submission to ActivityNet Challenge 2016

This paper presents the method that underlies our submission to the untrimmed video classification task of ActivityNet Challenge 2016. We follow the basic pipeline of temporal segment networks and further raise the performance via a number of other techniques. Specifically, we use the latest deep model architecture, e.g., ResNet and Inception V3, and introduce new aggregation schemes (top-k and attention-weighted pooling). Additionally, we incorporate the audio as a complementary channel, extracting relevant information via a CNN applied to the spectrograms. With these techniques, we derive an ensemble of deep models, which, together, attains a high classification accuracy (mAP $93.23\%$) on the testing set and secured the first place in the challenge.

cs.CV↗

arXiv2016

Temporal Segment Networks: Towards Good Practices for Deep Action Recognition

Deep convolutional networks have achieved great success for visual recognition in still images. However, for action recognition in videos, the advantage over traditional methods is not so evident. This paper aims to discover the principles to design effective ConvNet architectures for action recognition in videos and learn these models given limited training samples. Our first contribution is temporal segment network (TSN), a novel framework for video-based action recognition. which is based on the idea of long-range temporal structure modeling. It combines a sparse temporal sampling strategy and video-level supervision to enable efficient and effective learning using the whole action video. The other contribution is our study on a series of good practices in learning ConvNets on video data with the help of temporal segment network. Our approach obtains the state-the-of-art performance on the datasets of HMDB51 ( $ 69.4\% $) and UCF101 ($ 94.2\% $). We also visualize the learned ConvNet models, which qualitatively demonstrates the effectiveness of temporal segment network and the proposed good practices.

cs.CV↗

arXiv2016

Real-time Action Recognition with Enhanced Motion Vector CNNs

The deep two-stream architecture exhibited excellent performance on video based action recognition. The most computationally expensive step in this approach comes from the calculation of optical flow which prevents it to be real-time. This paper accelerates this architecture by replacing optical flow with motion vector which can be obtained directly from compressed videos without extra calculation. However, motion vector lacks fine structures, and contains noisy and inaccurate motion patterns, leading to the evident degradation of recognition performance. Our key insight for relieving this problem is that optical flow and motion vector are inherent correlated. Transferring the knowledge learned with optical flow CNN to motion vector CNN can significantly boost the performance of the latter. Specifically, we introduce three strategies for this, initialization transfer, supervision transfer and their combination. Experimental results show that our method achieves comparable recognition performance to the state-of-the-art, while our method can process 390.7 frames per second, which is 27 times faster than the original two-stream method.

cs.CV↗

arXiv2016

Actionness Estimation Using Hybrid Fully Convolutional Networks

Actionness was introduced to quantify the likelihood of containing a generic action instance at a specific location. Accurate and efficient estimation of actionness is important in video analysis and may benefit other relevant tasks such as action recognition and action detection. This paper presents a new deep architecture for actionness estimation, called hybrid fully convolutional network (H-FCN), which is composed of appearance FCN (A-FCN) and motion FCN (M-FCN). These two FCNs leverage the strong capacity of deep models to estimate actionness maps from the perspectives of static appearance and dynamic motion, respectively. In addition, the fully convolutional nature of H-FCN allows it to efficiently process videos with arbitrary sizes. Experiments are conducted on the challenging datasets of Stanford40, UCF Sports, and JHMDB to verify the effectiveness of H-FCN on actionness estimation, which demonstrate that our method achieves superior performance to previous ones. Moreover, we apply the estimated actionness maps on action proposal generation and action detection. Our actionness maps advance the current state-of-the-art performance of these tasks substantially.

cs.CV↗

arXiv2016

Pairing of j=3/2 fermions in half-Heusler superconductors

We theoretically consider the superconductivity of the topological half-Heusler semimetals YPtBi and LuPtBi. We show that pairing occurs between j=3/2 fermion states, which leads to qualitative differences from the conventional theory of pairing between j=1/2 states. In particular, this permits Cooper pairs with quintet or septet total angular momentum, in addition to the usual singlet and triplet states. Purely on-site interactions can generate s-wave quintet time-reversal symmetry-breaking states with topologically nontrivial point or line nodes. These local s-wave quintet pairs reveal themselves as d-wave states in momentum space. Furthermore, due to the broken inversion symmetry in these materials, the s-wave singlet state can mix with a p-wave septet state, again with topologically-stable line nodes. Our analysis lays the foundation for understanding the unconventional superconductivity of the half-Heuslers.

cond-mat.supr-con↗

arXiv2016

Enhanced Thermoelectric Power and Electronic Correlations in RuSe$_2$

We report the electronic structure, electric and thermal transport properties of Ru$_{1-x}$Ir$_{x}$Se$_2$ ($x \leq 0.2$). RuSe$_2$ is a semiconductor that crystallizes in a cubic pyrite unit cell. The Seebeck coefficient of RuSe$_2$ exceeds -200 $μ$V/K around 730 K. Ir substitution results in the suppression of the resistivity and the Seebeck coefficient, suggesting the removal of the peaks in density of states near the Fermi level. Ru$_{0.8}$Ir$_{0.2}$Se$_{2}$ shows a semiconductor-metal crossover at about 30 K. The magnetic field restores the semiconducting behavior. Our results indicate the importance of the electronic correlations in enhanced thermoelectricity of RuSb$_{2}$.

cond-mat.str-el↗

arXiv2016

Electron-hole asymmetry, Dirac fermions, and quantum magnetoresistance in BaMnBi2

We report two-dimensional quantum transport and Dirac fermions in BaMnBi2 single crystals. BaMnBi2 is a layered bad metal with highly anisotropic conductivity and magnetic order below 290 K. Magnetotransport properties, nonzero Berry phase, small cyclotronmass, and the first-principles band structure calculations indicate the presence of Dirac fermions in Bi square nets. Quantum oscillations in the Hall channel suggest the presence of both electron and hole pockets, whereas Dirac and parabolic states coexist at the Fermi level.

cond-mat.mtrl-sci↗

arXiv2015

Better Exploiting OS-CNNs for Better Event Recognition in Images

Event recognition from still images is one of the most important problems for image understanding. However, compared with object recognition and scene recognition, event recognition has received much less research attention in computer vision community. This paper addresses the problem of cultural event recognition in still images and focuses on applying deep learning methods on this problem. In particular, we utilize the successful architecture of Object-Scene Convolutional Neural Networks (OS-CNNs) to perform event recognition. OS-CNNs are composed of object nets and scene nets, which transfer the learned representations from the pre-trained models on large-scale object and scene recognition datasets, respectively. We propose four types of scenarios to explore OS-CNNs for event recognition by treating them as either "end-to-end event predictors" or "generic feature extractors". Our experimental results demonstrate that the global and local representations of OS-CNNs are complementary to each other. Finally, based on our investigation of OS-CNNs, we come up with a solution for the cultural event recognition track at the ICCV ChaLearn Looking at People (LAP) challenge 2015. Our team secures the third place at this challenge and our result is very close to the best performance.

cs.CV↗

arXiv2015

Places205-VGGNet Models for Scene Recognition

VGGNets have turned out to be effective for object recognition in still images. However, it is unable to yield good performance by directly adapting the VGGNet models trained on the ImageNet dataset for scene recognition. This report describes our implementation of training the VGGNets on the large-scale Places205 dataset. Specifically, we train three VGGNet models, namely VGGNet-11, VGGNet-13, and VGGNet-16, by using a Multi-GPU extension of Caffe toolbox with high computational efficiency. We verify the performance of trained Places205-VGGNet models on three datasets: MIT67, SUN397, and Places205. Our trained models achieve the state-of-the-art performance on these datasets and are made public available.

cs.CV↗

arXiv2015

Towards Good Practices for Very Deep Two-Stream ConvNets

Deep convolutional networks have achieved great success for object recognition in still images. However, for action recognition in videos, the improvement of deep convolutional networks is not so evident. We argue that there are two reasons that could probably explain this result. First the current network architectures (e.g. Two-stream ConvNets) are relatively shallow compared with those very deep models in image domain (e.g. VGGNet, GoogLeNet), and therefore their modeling capacity is constrained by their depth. Second, probably more importantly, the training dataset of action recognition is extremely small compared with the ImageNet dataset, and thus it will be easy to over-fit on the training dataset. To address these issues, this report presents very deep two-stream ConvNets for action recognition, by adapting recent very deep architectures into video domain. However, this extension is not easy as the size of action recognition is quite small. We design several good practices for the training of very deep two-stream ConvNets, namely (i) pre-training for both spatial and temporal nets, (ii) smaller learning rates, (iii) more data augmentation techniques, (iv) high drop out ratio. Meanwhile, we extend the Caffe toolbox into Multi-GPU implementation with high computational efficiency and low memory consumption. We verify the performance of very deep two-stream ConvNets on the dataset of UCF101 and it achieves the recognition accuracy of $91.4\%$.

cs.CV↗

arXiv2015

Action Recognition with Trajectory-Pooled Deep-Convolutional Descriptors

Visual features are of vital importance for human action understanding in videos. This paper presents a new video representation, called trajectory-pooled deep-convolutional descriptor (TDD), which shares the merits of both hand-crafted features and deep-learned features. Specifically, we utilize deep architectures to learn discriminative convolutional feature maps, and conduct trajectory-constrained pooling to aggregate these convolutional features into effective descriptors. To enhance the robustness of TDDs, we design two normalization methods to transform convolutional feature maps, namely spatiotemporal normalization and channel normalization. The advantages of our features come from (i) TDDs are automatically learned and contain high discriminative capacity compared with those hand-crafted features; (ii) TDDs take account of the intrinsic characteristics of temporal dimension and introduce the strategies of trajectory-constrained sampling and pooling for aggregating deep-learned features. We conduct experiments on two challenging datasets: HMDB51 and UCF101. Experimental results show that TDDs outperform previous hand-crafted features and deep-learned features. Our method also achieves superior performance to the state of the art on these datasets (HMDB51 65.9%, UCF101 91.5%).

cs.CV↗

arXiv2015

Object-Scene Convolutional Neural Networks for Event Recognition in Images

Event recognition from still images is of great importance for image understanding. However, compared with event recognition in videos, there are much fewer research works on event recognition in images. This paper addresses the issue of event recognition from images and proposes an effective method with deep neural networks. Specifically, we design a new architecture, called Object-Scene Convolutional Neural Network (OS-CNN). This architecture is decomposed into object net and scene net, which extract useful information for event understanding from the perspective of objects and scene context, respectively. Meanwhile, we investigate different network architectures for OS-CNN design, and adapt the deep (AlexNet) and very-deep (GoogLeNet) networks to the task of event recognition. Furthermore, we find that the deep and very-deep networks are complementary to each other. Finally, based on the proposed OS-CNN and comparative study of different network architectures, we come up with a solution of five-stream CNN for the track of cultural event recognition at the ChaLearn Looking at People (LAP) challenge 2015. Our method obtains the performance of 85.5% and ranks the $1^{st}$ place in this challenge.

cs.CV↗

arXiv2015

High-temperature superconductivity stabilized by electron-hole interband coupling in collapsed tetragonal phase of KFe2As2 under high pressure

We report a high-pressure study of simultaneous low-temperature electrical resistivity and Hall effect measurements on high quality single-crystalline KFe2As2 using designer diamond anvil cell techniques with applied pressures up to 33 GPa. In the low pressure regime, we show that the superconducting transition temperature T_c finds a maximum onset value of 7 K near 2 GPa, in contrast to previous reports that find a minimum T_c and reversal of pressure dependence at this pressure. Upon applying higher pressures, this T_c is diminished until a sudden drastic enhancement occurs coincident with a first-order structural phase transition into a collapsed tetragonal phase. The appearance of a distinct superconducting phase above 13 GPa is also accompanied by a sudden reversal of dominant charge carrier sign, from hole- to electron-like, which agrees with our band calculations predicting the emergence of an electron pocket and diminishment of hole pockets upon Fermi surface reconstruction. Our results suggest the high-temperature superconducting phase in KFe2As2 is substantially enhanced by the presence of nested electron and hole pockets, providing the key ingredient of high-T_c superconductivity in iron pnictide superconductors.

cond-mat.supr-con↗

arXiv2014

Quantum Walks of Two Interacting Anyons in 1D Optical Lattices

We investigate continuous-time quantum walks of two indistinguishable anyons in one-dimensional lattices with both on-site and nearest-neighbor interactions based on the fractional Jordan-Wigner transformation. It is shown that the two-body correlations in position space are symmetric about the initial sites of two quantum walkers in the Bose limit ($χ=0$ ) and Fermi limit ( $χ=1$), while in momentum space this happens only in the Bose limit. An interesting asymmetry arises in the correlation for most cases with the statistical parameter $χ$ varying in between. It turns out that the origin of this asymmetry comes from the fractional statistics that anyons obey. On the other hand, the two-body correlations of hard-core anyons in position spaceshow uniform behaviors from anti-bunching to co-walking regardless of the statistical parameter. The momentum correlations in the case of strong interaction undergo a smooth process of two stripes smoothly merging into a single one, i.e. the evolution of fermions into hard-core bosons.

cond-mat.quant-gas↗

arXiv2014

Lattice Boltzmann Model for The Volume-Averaged Navier-Stokes Equations

A numerical method, based on the discrete lattice Boltzmann equation, is presented for solving the volume-averaged Navier-Stokes equations. With a modified equilibrium distribution and an additional forcing term, the volume-averaged Navier-Stokes equations can be recovered from the lattice Boltzmann equation in the limit of small Mach number by the Chapman-Enskog analysis and Taylor expansion. Due to its advantages such as explicit solver and inherent parallelism, the method appears to be more competitive with traditional numerical techniques. Numerical simulations show that the proposed model can accurately reproduce both the linear and nonlinear drag effects of porosity in the fluid flow through porous media.

physics.flu-dyn↗

arXiv2014

Bag of Visual Words and Fusion Methods for Action Recognition: Comprehensive Study and Good Practice

Video based action recognition is one of the important and challenging problems in computer vision research. Bag of Visual Words model (BoVW) with local features has become the most popular method and obtained the state-of-the-art performance on several realistic datasets, such as the HMDB51, UCF50, and UCF101. BoVW is a general pipeline to construct a global representation from a set of local features, which is mainly composed of five steps: (i) feature extraction, (ii) feature pre-processing, (iii) codebook generation, (iv) feature encoding, and (v) pooling and normalization. Many efforts have been made in each step independently in different scenarios and their effect on action recognition is still unknown. Meanwhile, video data exhibits different views of visual pattern, such as static appearance and motion dynamics. Multiple descriptors are usually extracted to represent these different views. Many feature fusion methods have been developed in other areas and their influence on action recognition has never been investigated before. This paper aims to provide a comprehensive study of all steps in BoVW and different fusion methods, and uncover some good practice to produce a state-of-the-art action recognition system. Specifically, we explore two kinds of local features, ten kinds of encoding methods, eight kinds of pooling and normalization strategies, and three kinds of fusion methods. We conclude that every step is crucial for contributing to the final recognition rate. Furthermore, based on our comprehensive study, we propose a simple yet effective representation, called hybrid representation, by exploring the complementarity of different BoVW frameworks and local descriptors. Using this representation, we obtain the state-of-the-art on the three challenging datasets: HMDB51 (61.1%), UCF50 (92.3%), and UCF101 (87.9%).

cs.CV↗

arXiv2013

A stability condition for turbulence model: From EMMS model to EMMS-based turbulence model

The closure problem of turbulence is still a challenging issue in turbulence modeling. In this work, a stability condition is used to close turbulence. Specifically, we regard single-phase flow as a mixture of turbulent and non-turbulent fluids, separating the structure of turbulence. Subsequently, according to the picture of the turbulent eddy cascade, the energy contained in turbulent flow is decomposed into different parts and then quantified. A turbulence stability condition, similar to the principle of the energy-minimization multi-scale (EMMS) model for gas-solid systems, is formulated to close the dynamic constraint equations of turbulence, allowing the heterogeneous structural parameters of turbulence to be optimized. We call this model the `EMMS-based turbulence model', and use it to construct the corresponding turbulent viscosity coefficient. To validate the EMMS-based turbulence model, it is used to simulate two classical benchmark problems, lid-driven cavity flow and turbulent flow with forced convection in an empty room. The numerical results show that the EMMS-based turbulence model improves the accuracy of turbulence modeling due to it considers the principle of compromise in competition between viscosity and inertia.

physics.flu-dyn↗