The classification of heartbeats is an important method for cardiac arrhythmia analysis. This study proposes a novel heartbeat classification method using hybrid time-frequency analysis and transfer learning based on ResNet-101. The proposed method has the following major advantages over the afore-mentioned methods: it avoids the need for manual features extraction in the traditional machine learning method, and it utilizes 2-D time-frequency diagrams which provide not only frequency and energy information but also preserve the morphological characteristic within the ECG recordings, and it owns enough deep to make better use of performance of CNN. The method deploys a hybrid time-frequency analysis of the Hilbert transform (HT) and the Wigner-Ville distribution (WVD) to transform 1-D ECG recordings into 2-D time-frequency diagrams which were then fed into a transfer learning classifier based on ResNet-101 for two classification tasks (i.e., 5 heartbeat categories assigned by the ANSI/AAMI standard (i.e., N, V, S, Q and F) and 14 original beat kinds of the MIT/BIH arrhythmia database). For 5 heartbeat categories classification, the results show the F1-score of N, V, S, Q and F categories are F$_{N}$ 0.9899, F$_{V}$ 0.9845, F$_{S}$ 0.9376, F$_{Q}$ 0.9968, F$_{F}$ 0.8889, respectively, and the overall F1-score is 0.9595 using the combination data balancing. The results show the average values for accuracy, sensitivity, specificity, predictive value and F1-score on test set for 14 beat kinds the MIT-BIH arrhythmia database are 99.75%, 91.36%, 99.85%, 90.81% and 0.9016, respectively. Compared with other methods, the proposed method can yield more accurate results.

Complex network models are helpful to explain the evolution rules of network structures, and also are the foundations of understanding and controlling complex networks. The existing studies (e.g., scale-free model, small-world model) are insufficient to uncover the internal mechanisms of the emergence and evolution of communities in networks. To overcome the above limitation, in consideration of the fact that a network can be regarded as a pattern composed of communities, we introduce Turing pattern dynamic as theory support to construct the network evolution model. Specifically, we develop a Reaction-Diffusion model according to Q-Learning technology (RDQL), in which each node regarded as an intelligent agent makes a behavior choice to update its relationships, based on the utility and behavioral strategy at every time step. Extensive experiments indicate that our model not only reveals how communities form and evolve, but also can generate networks with the properties of scale-free, small-world and assortativity. The effectiveness of the RDQL model has also been verified by its application in real networks. Furthermore, the depth analysis of the RDQL model provides a conclusion that the proportion of exploration and exploitation behaviors of nodes is the only factor affecting the formation of communities. The proposed RDQL model has potential to be the basic theoretical tool for studying network stability and dynamics.

The network topology of complex networks evolves dynamically with time. How to model the internal mechanism driving the dynamic change of network structure is the key problem in the field of complex networks. The models represented by WS, NW, BA usually assume that the evolution of network structure is driven by nodes’ passive behaviors based on some restrictive rules. However, in fact, network nodes are intelligent individuals, which actively update their relations based on experience and environment. To overcome this limitation, we attempt to construct a network model based on deep reinforcement learning, named as NMDRL. In the new model, each node in complex networks is regarded as an intelligent agent, which reacts with the agents around it for refreshing its relationships at every moment. Extensive experiments show that our model not only can generate networks owing the properties of scale-free and small-world, but also reveal how community structures emerge and evolve. The proposed NMDRL model is helpful to study propagation, game, and cooperation behaviors in networks.

Influence maximization is a hot research topic in the social computing field and has gained tremendous studies motivated by its wild application scenarios. As the structures of social networks change over time, how to seek seed node sets from dynamic networks has attracted some attention. However, all of the existing studies were based on network topology structure data which have the limitations of high dimensionality and low efficiency. Aiming at this drawback, we first convert each node in the network to a low-dimensional vector representation by network representation learning and then solve the problem of dynamic influence maximization in the low-dimensional latent space. Comprehensive experiments on NetHEPT, Twitter, UCI, and Wikipedia datasets show that our method can achieve influence diffusion performance similar to state-of-the-art approaches in much less time.

Image super-resolution (SR) techniques aim to enhance the clarity and realism of images. Recently, a wide range of excellent SR algorithms with powerful characterization capabilities have emerged and are widely used. However, there are still challenges and room for improvement in designing a lighter and more edge-friendly SR networks for hardware devices. In this paper, we propose a lightweight enhanced feature refinement network (EFRN) based on depthwise separatable convolution for SR reconstruction. The core network components consist of multiple enhanced feature refinement blocks (EFRB), which fully fuse channel features to extract more accurate low-frequency information based on the attention of different channels. In addition, a lightweight residual block (LRB) and a lightweight dual attention block (LDAB) are designed to enhance network information extraction with minimal parameter cost. We improve the feature refinement by using 1 × 1 convolution instead of a channel selection operation to reduce the dimensionality of the features and extract the refined features more efficiently. Finally, to achieve better reconstruction performance, the depth and number of channels of the network are expanded while keeping the total number of parameters at a low level. Extensive experiments have been conducted to demonstrate the superiority of our EFRN over other mainstream SR algorithms in terms of reconstruction results and the number of parameters.

We present a novel approach to non-rigid object tracking in this paper by deriving an adaptive data-driven kernel. In contrast with conventional kernel-based trackers which suffer from the constancy of kernel shape as well as scale and orientation selection problem when the tracking targets are changing in size, the adaptive kernel can robustly achieve the adaptation to target variation and act toward the actual target contour simultaneously with the mean shift iterations. Level set technique is novelly introduced to the mean shift sample space to both cope with insufficient low-level information and implement the adaptive kernel evolution and update. Since the active contour model is designed to drive the kernel constantly to the direction that maximizes the appearance similarity, this adaptive kernel can continually seize the target shape to give a better estimation bias and produce accurate shift of the mean. Finally, accurate target region can successfully avoid the performance loss stemmed from pollution of background pixels hiding inside the kernel and qualify the samples fed the next time step. Experimental results on a numer of challenging sequences validate the effectiveness of the technique.