• JSTS:Journal of Semiconductor Technology and Science
• /
• v.13 no.1
• /
• pp.8-14
• /
• 2013

A low density parity check (LDPC) decoder provides a most powerful error control capability for mobile communication devices and storage systems, due to its performance being close to Shannon's limit. In this paper, we introduce an efficient overlapped LDPC decoding algorithm using a upper dual-diagonal parity check matrix structure. By means of this algorithm, the LDPC decoder can concurrently execute parts of the check node update and variable node update in the sum-product algorithm. In this way, we can reduce the number of clock cycles per iteration as well as reduce the total latency. The proposed decoding structure offers a very simple control and is very flexible in terms of the variable bit length and variable code rate. The experiment results show that the proposed decoder can complete the decoding of codewords within 70% of the number of clock cycles required for a conventional non-overlapped decoder. The proposed design also reduces the power consumption by 33% when compared to the non-overlapped design.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.35C no.10
• /
• pp.48-58
• /
• 1998

We present a method that can recognize face images using independent component analysis that is used mainly for blind sources separation in signal processing. We assumed that a face image can be expressed as the sum of a set of statistically independent feature images, which was obtained by using independent component analysis. Face recognition was peformed by projecting the input image to the feature image space and then by comparing its projection components with those of stored reference images. We carried out face recognition experiments with a database that consists of various varied face images (total 400 varied facial images collected from 10 per person) and compared the performance of our method with that of the eigenface method based on principal component analysis. The presented method gave better results of recognition rate than the eigenface method did, and showed robustness to the random noise added in the input facial images.

• Phonetics and Speech Sciences
• /
• v.7 no.3
• /
• pp.131-138
• /
• 2015

In this paper, we propose L1-norm regularization for state vector adaptation of subspace Gaussian mixture model (SGMM). When you design a speaker adaptation system with GMM-HMM acoustic model, MAP is the most typical technique to be considered. However, in MAP adaptation procedure, large number of parameters should be updated simultaneously. We can adopt sparse adaptation such as L1-norm regularization or sparse MAP to cope with that, but the performance of sparse adaptation is not good as MAP adaptation. However, SGMM does not suffer a lot from sparse adaptation as GMM-HMM because each Gaussian mean vector in SGMM is defined as a weighted sum of basis vectors, which is much robust to the fluctuation of parameters. Since there are only a few adaptation techniques appropriate for SGMM, our proposed method could be powerful especially when the number of adaptation data is limited. Experimental results show that error reduction rate of the proposed method is better than the result of MAP adaptation of SGMM, even with small adaptation data.

• Journal of Communications and Networks
• /
• v.15 no.1
• /
• pp.102-110
• /
• 2013

This paper considers a cooperative orthogonal frequency division multiple access (OFDMA)-based cognitive radio network where the primary system leases some of its subchannels to the secondary system for a fraction of time in exchange for the secondary users (SUs) assisting the transmission of primary users (PUs) as relays. Our aim is to determine the cooperation strategies among the primary and secondary systems so as to maximize the sum-rate of SUs while maintaining quality-of-service (QoS) requirements of PUs. We formulate a joint optimization problem of PU transmission mode selection, SU (or relay) selection, subcarrier assignment, power control, and time allocation. By applying dual method, this mixed integer programming problem is decomposed into parallel per-subcarrier subproblems, with each determining the cooperation strategy between one PU and one SU. We show that, on each leased subcarrier, the optimal strategy is to let a SU exclusively act as a relay or transmit for itself. This result is fundamentally different from the conventional spectrum leasing in single-channel systems where a SU must transmit a fraction of time for itself if it helps the PU's transmission. We then propose a subgradient-based algorithm to find the asymptotically optimal solution to the primal problem in polynomial time. Simulation results demonstrate that the proposed algorithm can significantly enhance the network performance.

• Journal of the Korean Society of Safety
• /
• v.26 no.3
• /
• pp.83-90
• /
• 2011

It is well known that urgency resulted from time stress can be a great cause to industrial accidents. Therefore, time stress has been studied in the aspect of macroscopic view, namely industrial safety management, but has not been studied in microscopic view such as psychophysiological approach. Among diverse psychophysiological indices, Electroencephalogram(EEG) would be on of the most objective psychophysiological research technique on human errors though few research has been taken yet. This study aimed to get characteristics of human error while committing a simple arithmetic addition task by utilizing the power spectrum technique of EEG data. Each experiment was composed of 2 tasks under different condition - with and without time stress. As subjects, 5 young undergraduate students in their early twenties participated in this study. The results advocated a well-known fact that time stress downgrades the performance of human workers. However, correct answer rate and response time were not significantly influenced by time stress factor which might be explained by the constructural factor adopted in the present study. As in the previous studies, among various EEG-related measures, relative band power ratios of ${\alpha}$ and ${\beta}$ waves to sum of ${\alpha}$,${\beta}$,${\theta}$ wave powers, namely $P_{{\alpha}/({\alpha}+{\beta}+{\theta})}$ and $P_{{\beta}/({\alpha}+{\beta}+{\theta})}$ seemed to be the most effective measures to grasp variation of brain activities in time-stressed situation so that discussions were expanded about their variations.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.11
• /
• pp.73-79
• /
• 2009

In this paper, a method is proposed that enhances the performance of the facial expression recognition using template matching of Local Binary Pattern(LBP) histogram. In this method, the face image is segmented into blocks, and the LBP histogram is constructed to be used as the feature of the block. Block dissimilarity is calculated between a block of input image and the corresponding block of the model image. Image dissimilarity is defined as the weighted sum of the block dissimilarities. In conventional methods, the block weights are assigned by intuition. In this paper a new method is proposed that optimizes the weights from training samples. An experiment shows the recognition rate is enhanced by the proposed method.

• ETRI Journal
• /
• v.45 no.6
• /
• pp.963-973
• /
• 2023

Non-orthogonal multiple access (NOMA) is considered a key candidate technology for next-generation wireless communication systems due to its high spectral efficiency and massive connectivity. Incorporating the concepts of multiple-input-multiple-output (MIMO) into NOMA can further improve the system efficiency, but the hardware complexity increases. This study develops an energy-efficient (EE) subchannel assignment framework for MIMO-NOMA systems under the quality-of-service and interference constraints. This framework handles an energy-efficient co-training-based semi-supervised learning (EE-CSL) algorithm, which utilizes a small portion of existing labeled data generated by numerical iterative algorithms for training. To improve the learning performance of the proposed EE-CSL, initial assignment is performed by a many-to-one matching (MOM) algorithm. The MOM algorithm helps achieve a low complex solution. Simulation results illustrate that a lower computational complexity of the EE-CSL algorithm helps significantly minimize the energy consumption in a network. Furthermore, the sum rate of NOMA outperforms conventional orthogonal multiple access.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.23 no.5
• /
• pp.1325-1341
• /
• 1998

A conventional CBR channel is now capable of delivering several digitally compressed video programs due to recent advances in video compression, such as MPEG-2, and digital transmission technology. This paper presents a joint video coding scheme that is to maintain a constant sum of bit rates for all the programs but to allow the variable bit rate for individual program in the transimission environment mentioned above. Thus advantages of VBR video compression can be obtained. This paper contributes in two aspects. First, a rate-distortion estimation method for MPEG-2 video is proposed, which enavle us predict the amount of bits and the distortion generated from an encoded picture at a given quantization step size and vice versa. The most attractive features of the proposed rate-distortion estimation method are its accuracy and a computational complexity low enough to be applied to real-time video coding applications. Second, this paper presents an efficient and accurate joint rate control scheme using the rate-distortion estimation results and verifies its performance with experiments. The experimental results show that our coding scheme gives a significant gain even though a small number of video programs are coded jointly. For example, a stable picture quality is maintained among the video programs as well as within a program, and additional extra programs can be transmitted over the same channel bandwidth if the proposed joint video coding scheme is used.

• The KIPS Transactions:PartA
• /
• v.19A no.1
• /
• pp.35-44
• /
• 2012

Traditional technologies that are used to improve the performance of hard disk drives show many negative cases if they are applied to solid state drives (SSD). Access time and block sequence in hard disk drives that consist of mechanical components are very important performance factors. Meanwhile, SSD provides superior random read performance that is not affected by block address sequence due to the characteristics of flash memory. Practically, it is recommended to disable prefetching if a SSD is installed in a personal computer. However, this paper presents a combinational method of a prefetching scheme and a memory management that consider the internal structure of SSD and the characteristics of NAND flash memory. It is important that SSD must concurrently operate multiple flash memory chips. The I/O unit size of NAND flash memory tends to increase and it exceeded the block size of operating systems. Hence, the proposed prefetching scheme performs in an operating unit of SSD. To complement a weak point of the prefetching scheme, the proposed memory management scheme adaptively evicts uselessly prefetched data to maximize the sum of cache hit rate and prefetch hit rate. We implemented the proposed schemes as a Linux kernel module and evaluated them using a commercial SSD. The schemes improved the I/O performance up to 26% in a given experiment.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.9
• /
• pp.11-19
• /
• 2009

A RAM-based Neural Net(RBNN) which has multi-discriminators is more effective than RBNN with a discriminator. Experience Sensitive Cumulative Neural Network and 3-D Neuro System(3DNS) that accumulate the features point improved the performance of BNN, which were enabled to train additional and repeated patterns and extract a generalized pattern. In recognition process of Neural Net with multi-discriminator, the selection of class was decided by the value of MRD which calculates the accumulated sum of each class. But they had a saturation problem of its memory cells caused by learning volume increment. Therefore, the decision of MRD has a low performance because recognition rate is decreased by saturation. In this paper, we propose the method which improve the MRD ability. The method consists of the optimum MRD and the matching ratio prototype to generalized image, the cumulative filter ratio, the gap of prototype response MRD. We experimented the performance using NIST database of NIST without preprocessor, and compared this model with 3DNS. The proposed MRD method has more performance of recognition rate and more stable system for distortion of input pattern than 3DNS.