• Proceedings of the Korean Magnestics Society Conference
• /
• 2003.12a
• /
• pp.91-91
• /
• 2003

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.12
• /
• pp.6080-6096
• /
• 2019

Text classification is one of the fundamental techniques in natural language processing. Numerous studies are based on text classification, such as news subject classification, question answering system classification, and movie review classification. Traditional text classification methods are used to extract features and then classify them. However, traditional methods are too complex to operate, and their accuracy is not sufficiently high. Recently, convolutional neural network (CNN) based one-hot method has been proposed in text classification to solve this problem. In this paper, we propose an improved method using CNN based skip-gram method for Chinese text classification and it conducts in Sogou news corpus. Experimental results indicate that CNN with the skip-gram model performs more efficiently than CNN-based one-hot method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.9
• /
• pp.4549-4566
• /
• 2017

This work presents a novel facial descriptor, which is named as multiscale adaptive local directional texture pattern (MALDTP) and employed for expression recognition. We apply an adaptive threshold value to encode facial image in different scales, and concatenate a series of histograms based on the MALDTP to generate facial descriptor in term of Gabor filters. In addition, some dedicated experiments were conducted to evaluate the performance of the MALDTP method in a person-independent way. The experimental results demonstrate that our proposed method achieves higher recognition rate than local directional texture pattern (LDTP). Moreover, the MALDTP method has lower computational complexity, fewer storage space and higher classification accuracy than local Gabor binary pattern histogram sequence (LGBPHS) method. In a nutshell, the proposed MALDTP method can not only avoid choosing the threshold by experience but also contain much more structural and contrast information of facial image than LDTP.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.6
• /
• pp.2450-2469
• /
• 2018

In this paper, we develop a spatiotemporal model to analysis of cellular user in underlay D2D communication by using stochastic geometry and queuing theory. Firstly, by exploring stochastic geometry to model the user locations, we derive the probability that the SINR of cellular user in a predefined interval, which constrains the corresponding transmission rate of cellular user. Secondly, in contrast to the previous studies with full traffic models, we employ queueing theory to evaluate the performance parameters of dynamic traffic model and formulate the cellular user transmission mechanism as a M/G/1 queuing model. In the derivation, Embedded Markov chain is introduced to depict the stationary distribution of cellular user queue status. Thirdly, the expressions of performance metrics in terms of mean queue length, mean throughput, mean delay and mean dropping probability are obtained, respectively. Simulation results show the validity and rationality of the theoretical analysis under different channel conditions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.5
• /
• pp.1654-1675
• /
• 2014

In practical wireless systems, the erroneous channel state information (CSI) sometimes deteriorates the performance drastically. This paper focuses on robust design of coordinated set planning of coordinated multi-point (CoMP) transmission, with respect to the feedback delay and link error. The non-ideal channel models involving various uncertainty conditions are given. After defining a penalty factor, the robust net ergodic capacity optimization problem is derived, whose variables to be optimized are the number of coordinated base stations (BSs) and the divided area's radius. By the maximum minimum criterion, upper and lower bounds of the robust capacity are investigated. A practical scheme is proposed to determine the optimal number of cooperative BSs. The simulation results indicate that the robust design based on maxmin principle is better than other precoding schemes. The gap between two bounds gets smaller as transmission power increases. Besides, as the large scale fading is higher or the channel is less reliable, the number of the cooperated BSs shall be greater.

• Proceedings of the Korean Magnestics Society Conference
• /
• 2003.12a
• /
• pp.109-109
• /
• 2003

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.10
• /
• pp.3887-3907
• /
• 2015

This paper introduces a full duplex single secondary user multiple-input multiple-output (FD-SSU-MIMO) cognitive radio network, where secondary user (SU) opportunistically accesses the authorized spectrum unoccupied by primary user (PU) and transmits data based on FD-MIMO mode. Then we study the network achievable average sum-rate maximization problem under sum transmit power budget constraint at SU communication nodes. In order to solve the trade-off problem between SU's sensing time and data transmission time based on opportunistic spectrum access (OSA) and the power allocation problem based on FD-MIMO transmit mode, we propose a simple trisection algorithm to obtain the optimal sensing time and apply an alternating optimization (AO) algorithm to tackle the FD-MIMO based network achievable sum-rate maximization problem. Simulation results show that our proposed sensing time optimization and AO-based optimal power allocation strategies obtain a higher achievable average sum-rate than sequential convex approximations for matrix-variable programming (SCAMP)-based power allocation for the FD transmission mode, as well as equal power allocation for the half duplex (HD) transmission mode.

• ETRI Journal
• /
• v.44 no.3
• /
• pp.462-475
• /
• 2022

Although researchers have proposed numerous techniques for speech emotion recognition, its performance remains unsatisfactory in many application scenarios. In this study, we propose a speech emotion recognition model based on a genetic algorithm (GA)-decision tree (DT) fusion of deep and acoustic features. To more comprehensively express speech emotional information, first, frame-level deep and acoustic features are extracted from a speech signal. Next, five kinds of statistic variables of these features are calculated to obtain utterance-level features. The Fisher feature selection criterion is employed to select high-performance features, removing redundant information. In the feature fusion stage, the GA is is used to adaptively search for the best feature fusion weight. Finally, using the fused feature, the proposed speech emotion recognition model based on a DT support vector machine model is realized. Experimental results on the Berlin speech emotion database and the Chinese emotion speech database indicate that the proposed model outperforms an average weight fusion method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.2
• /
• pp.810-831
• /
• 2019

Recently, continuous dimensional emotion recognition from audiovisual clues has attracted increasing attention in both theory and in practice. The large amount of data involved in the recognition processing decreases the efficiency of most bimodal information fusion algorithms. A novel algorithm, namely the incomplete Cholesky decomposition based kernel cross factor analysis (ICDKCFA), is presented and employed for continuous dimensional audiovisual emotion recognition, in this paper. After the ICDKCFA feature transformation, two basic fusion strategies, namely feature-level fusion and decision-level fusion, are explored to combine the transformed visual and audio features for emotion recognition. Finally, extensive experiments are conducted to evaluate the ICDKCFA approach on the AVEC 2016 Multimodal Affect Recognition Sub-Challenge dataset. The experimental results show that the ICDKCFA method has a higher speed than the original kernel cross factor analysis with the comparable performance. Moreover, the ICDKCFA method achieves a better performance than other common information fusion methods, such as the Canonical correlation analysis, kernel canonical correlation analysis and cross-modal factor analysis based fusion methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.5
• /
• pp.2400-2413
• /
• 2019

Doubly-selective (DS) fading channel is often occurred in many orthogonal frequency division multiplexing (OFDM) communication systems, such as high-speed rail communication systems and underwater acoustic (UWA) wireless networks. It is challenging to provide an accurate and fast estimation over the doubly-selective channel, due to the strong Doppler shift. This paper addresses the doubly selective channel estimation problem based on complex exponential basis expansion model (CE-BEM) in OFDM systems from the perspective of distributed compressive sensing (DCS). We propose a novel DCS-based improved sparsity adaptive matching pursuit (DCS-IMSAMP) algorithm. The advantage of the proposed algorithm is that it can exploit the joint channel sparsity information using dynamic threshold, variable step size and tailoring mechanism. Simulation results show that the proposed algorithm achieves 5dB performance gain with faster operation speed, in comparison with traditional DCS-based sparsity adaptive matching pursuit (DCS-SAMP) algorithm.