Facial Expression Recognition Using Enhanced Convolution Neural Network with Attention Mechanism

Facial Expression Recognition(FER)has been an interesting area of research in places where there is human-computer interaction.Human psychol-ogy,emotions and behaviors can be analyzed in FER.Classifiers used in FER have been perfect on normal faces but have been found to be constrained in occluded faces.Recently,Deep Learning Techniques(DLT)have gained popular-ity in applications of real-world problems including recognition of human emo-tions.The human face reflects emotional states and human intentions.An expression is the most natural and powerful way of communicating non-verbally.Systems which form communications between the two are termed Human Machine Interaction(HMI)systems.FER can improve HMI systems as human expressions convey useful information to an observer.This paper proposes a FER scheme called EECNN(Enhanced Convolution Neural Network with Atten-tion mechanism)to recognize seven types of human emotions with satisfying results in its experiments.Proposed EECNN achieved 89.8%accuracy in classi-fying the images.

DIRECT DISPLACEMENT OF PARALLEL MECHANISM WITH WAVELET NETWORK

A new method solution for the direct displacement of parallel mechanism, wavelet network method, is proposed. Comparing with the classical analytical and numerical methods, this method can be extended to any parallel mechanism with any selected degree of freedom and configuration. A wavelet network suiting to approach multi-input and multi-output system is constructed. The network is optimized by analyzing the sparseness of input data and selecting the fitting wavelets by orthogonalization method according to the output data. Then it is applied to solve the direct displace- ment of a general six-degree-of-freedom parallel mechanism as a numerical example. For comparison purposes, a BP neural network is also used for this problem. Simulation results show that the wavelet network performs better than BP neural network. In addition, the wavelet network learns much faster than BP network.

Use of Network Pharmacology and Molecular Docking to Investigate the Mechanism by Which Ginseng Ameliorates Hypoxia

Hypoxia is a common pathological process in various clinical diseases and is characterized by abnormal changes in metabolism, function, and morphological structure of tissues resulting from insufficient oxygen supply or oxygen barriers in tissues. In particular, hypoxia in vital organs such as the brain and heart is an important cause of death;. The prevention of tissue hypoxia and the

Considerations on Novel Network Routing Mechanism

Routing is the cornerstone of network architecture, and a new routing mechanism is a requisite for constructing new network architecture. The current routing mechanism of the Internet layer is basically the single next-hop routing mechanism, which is the root of transmission congestion in the network. To solve this congestion problem, a fundamental measure is to change the route selection mode of current routing mechanism to allow parallel transmission over multiple routes. Currently, Border Gateway Protocol (BGP) is the only inter-domain routing protocol used in the Internet, but the routing system using BGP suffers scalability problem. To solve the scalability issue of the current inter-domain routing system, a new hierarchical routing system, i.e. Scalable Inter-Domain Routing Architecture (s-idra), is suggested. In addition to scalability, the current routing system is facing other challenges including security, Quality of Service (QoS), multicasting, mobility and dynamic network topology. Therefore, the research on routing protocols, especially the protocol for the new network architecture, is still a tough task and has a long way to go.

Fusion network for small target detection based on YOLO and attention mechanism

Target detection is an important task in computer vision research, and such an anomaly detection and the topic of small target detection task is more concerned. However, there are still some problems in this kind of researches, such as small target detection in complex environments is susceptible to background interference and poor detection results. To solve these issues, this study proposes a method which introduces the attention mechanism into the you only look once(YOLO) network. In addition, the amateur-produced mask dataset was created and experiments were conducted. The results showed that the detection effect of the proposed mothed is much better.

Traffic of indistinguishable particles in complex networks

In this paper, we apply a simple walk mechanism to the study of the traffic of many indistinguishable particles in complex networks. The network with particles stands for a particle system, and every vertex in the network stands for a quantum state with the corresponding energy determined by the vertex degree. Although the particles are indistinguishable, the quantum states can be distinguished. When the many indistinguishable particles walk randomly in the system for a long enough time and the system reaches dynamic equilibrium, we find that under different restrictive conditions the particle distributions satisfy different forms, including the Bose Einstein distribution, the Fermi Dirac distribution and the non-Fermi distribution （as we temporarily call it）. As for the Bose-Einstein distribution, we find that only if the particle density is larger than zero, with increasing particle density, do more and more particles condense in the lowest energy level. While the particle density is very low, the particle distribution transforms from the quantum statistical form to the classically statistical form, i.e., transforms from the Bose distribution or the Fermi distribution to the Boltzmann distribution. The numerical results fit well with the analytical predictions.

Molecular Mechanisms Regulating Rapid Stress Signaling Networks in Arabidopsis

As sessile organisms plants must ronmental conditions. To survive cope with ever changing enviplants have evolved elaborate mechanisms to perceive and rapidly respond to a diverse range of abiotic and biotic stresses. Central to this response is the ability to modulate gene expression at both the transcriptional and posttranscriptional levels. This review will focus on recent progress that has been made towards understanding the rapid reprogramming of the transcriptome that occurs in response to stress as well as emerging mechanisms underpinning the reprogramming of gene expression in response to stress,

Hybrid approach for loss recovery mechanism in OBS networks

This letter reports a study of a hybrid burst assembly and a hybrid burst loss recovery scheme （delay-based burst assembly and hybrid loss recovery （DBAHLR）） which selectively employs proactive or reactive loss recovery techniques depending on the classification of traffic into short term and long term, respectively. Traffic prediction and segregation of optical burst switching network flows into the long term and short term are conducted based on predicted link holding times using the hidden Markov model （HMM）. The hybrid burst assembly implemented in DBAHLR uses a consecutive average-based burst assembly to handle jitter reduction necessary in real-time applications, with variations in burst sizes due to the non-monotonic nature of the average delay handled by additional burst length thresholding. This dynamic hybrid approach based on HMM prediction provides overall a lower blocking probability and delay and more throughput when compared with forward segment redundancy mechanism or purely HMM prediction-based adaptive burst sizing and wavelength allocation （HMM-TP）.

Network Coding-Aware Routing Protocol in Wireless Mesh Networks

Network coding mechanisms, such as COPE, can improve network throughput effectively in Wireless Mesh Networks（WMN）. While the Hybrid Wireless Mesh Protocol（HWMP） is suitable for WMN, its extension with COPE does not provide any added benefits; specifically, HWMP cannot establish paths with more coding opportunities. As a result, the advantages of network coding cannot be exploited sufficiently. This paper proposes improvements upon HWMP with a new, network Coding-Aware routing protocol（CAHWMP） for WMN. In the CAHWMP protocol, we propose a coding criterion based on data streams to devise an algorithm for actively detecting coding opportunities during path discovery. CAHWMP subsequently establishes paths using the codingaware routing metric, which can balance channel resource consumption and the gain due to sharing resources introduced by network coding. Simulation results show that CAHWMP can establish paths with more coding opportunities; as a result, it improves network performance such as network throughput.

Insular Cortex is Critical for the Perception, Modulation,and Chronification of Pain

An increasing body of neuroimaging and electrophysiological studies of the brain suggest that the insular cortex（IC） integrates multimodal salient information ranging from sensation to cognitive-affective events to create conscious interoception. Especially with regard to pain experience, the IC has been supposed to participate in both sensory-discriminative and affective-motivational aspects of pain. In this review, we discuss the latest data proposing that subregions of the IC are involved in isolated pain networks： the posterior sensory circuit and the anterior emotional network. Due to abundant connections with other brain areas, the IC is likely to serve as an interface where cross-modal shaping of pain occurs. In chronic pain,however, this mode of emotional awareness and the modulation of pain are disrupted. We highlight some of the molecular mechanisms underlying the changes of the pain modulation system that contribute to the transition from acute to chronic pain in the IC.

Influences of the Three Gorges Project on seismic activities in the reservoir area

Reservoir-induced earthquakes related with the construction of the Three Gorges Project have attracted great concerns of the public. Since the first water impoundment on May 25, 2003, a number of earthquakes have occurred during the water storage stages, in which the largest was the Badong M5.1 earthquake on December 16, 2013. In this paper, the relationships between seismic activities, b value, seismic parameters, and reservoir water level fluctuations are studied. In addition, based on the digital seismic waveform data obtained since 2000, the focal depth changes and focal mechanism characteristics before and after the water impoundment are studied as well. These provide us important information to understand the earthquake mechanisms. The results show that these earthquakes are typical reservoir-induced earthquakes, which are closely related to water infiltration, pore pressure, and water level fluctuations.The majority of the micro and small earthquakes are caused by karst collapse, mine collapse, bank reformation, superficial unloading, and so on. The larger earthquakes are related to the fault structures to some extent. Due to the persistent effects of water impoundment on the seismic and geological environments around the reservoir and water infiltration into the rocks, the influences on the crustal deformation field, gravity field, seepage field, and fault medium-softening action may vary gradually from a higher strength to a weaker one. Therefore, it is possible that small earthquakes and few medium earthquakes(M≤5.5) will occur in the reservoir area in the future.