• Korean Journal of Artificial Intelligence
• /
• v.5 no.1
• /
• pp.10-17
• /
• 2017

This study is the first attempt to raise face recognition ability through machine learning algorithm and apply to CRM's information gathering, analysis and application. In other words, through face recognition of VIP customer in distribution field, we can proceed more prompt and subdivided customized services. The interest in machine learning, which is used to implement artificial intelligence, has increased, and it has become an age to automate it by using machine learning beyond the way that a person directly models an object recognition process. Among them, Deep Learning is evaluated as an advanced technology that shows amazing performance in various fields, and is applied to various fields of image recognition. Face recognition, which is widely used in real life, has been developed to recognize criminals' faces and catch criminals. In this study, two image analysis models, TF-SLIM and Inception-V3, which are likely to be used for criminal face recognition, were selected, analyzed, and implemented. As an evaluation criterion, the image recognition model was evaluated based on the accuracy of the face recognition program which is already being commercialized. In this experiment, it was evaluated that the recognition accuracy was good when the accuracy of the image classification was more than 90%. A limit of our study which is a way to raise face recognition is left as a further research subjects.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.4
• /
• pp.921-929
• /
• 2011

This paper describes an analysis of the decoding performance and decoding convergence speed of layered LDPC(low-density parity-check) decoder for mobile WiMAX system, and the optimal design conditions for hardware implementation are searched. A fixed-point model of LDPC decoder, which is based on the min-sum algorithm and layered decoding scheme, is implemented and simulated using Matlab model. Through fixed-point simulations for the block lengths of 576, 1440, 2304 bits and the code rates of 1/2, 2/3A, 2/3B, 3/4A, 3/4B, 5/6 specified in the IEEE 802.16e standard, the effect of internal bit-width, block length and code rate on the decoding performance are analyzed. Simulation results show that fixed-point bit-width larger than 8 bits with integer part of 5 bits should be used for acceptable decoding performance.

• The KIPS Transactions:PartA
• /
• v.12A no.3 s.93
• /
• pp.215-222
• /
• 2005

The efficient utilization of cache memories is a key factor in achieving high performance for computing large signal transforms. Nonunit stride access in computation of large DFTs causes cache conflict misses, thereby resulting in poor cache performance. It leads to a severe degradation in overall performance. In this paper, we propose a dynamic data layout approach considering the memory hierarchy system. In our approach, data reorganization is performed between computation stages to reduce the number of cache misses. Also, we develop an efficient search algorithm to determine the optimal tree with the minimum execution time among possible factorization trees considering the size of DFTs and the data access stride. Our approach is applied to compute the fast Fourier Transform (FFT). Experiments were performed on Pentium 4, $Athlon^{TM}$ 64, Alpha 21264, UtraSPARC III. Experiment results show that our FFT achieve performance improvement of up to 3.37 times better than the previous FFT packages.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38B no.1
• /
• pp.18-25
• /
• 2013

In this paper, we propose a self-optimization scheme for indoor femtocell coverage in heterogeneous networks. If the femtocell coverage is larger than indoor area, neighbor cell users passing by the outer area of the femtocell coverage may request an unnecessary handover which incurs wasteful signaling overhead. On the other hand, if the coverage is smaller than the indoor area, some of indoor users might not be connected to the indoor femtocell. Therefore, we propose the method by which the femtocell coverage attains the exact indoor area employing self-organized scheme. Autonomous self TX power adjustment of the femtocell is possible because the proposed method utilizes handover request events and membership information of users that can be obtained by the femtocell itself. We show that the TX power obtained by the proposed algorithm converges to the optimal TX power that can be obtained analytically to attain the indoor coverage area.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.15 no.9
• /
• pp.745-752
• /
• 1990

This paper deals with the path sensitization algrithm from PI to PO center on the nodes of high testability mainstay when CUT is tested by pseudo exhaustive testing. In CUT, the node definition of high testability mainstay treats the testability values of the entire nodes with the population composed of the raw data, and after we examined the level of significance(1-a) region, we accomplished in the estimation of the confidence interval of the testability. Focusing on the defined nodes of high testability mainstay, we performed the singular cover and consistency operation to the forward and backward logic gates. Thus, we easily generated the pseudo exhausitve test patterns. As a result, (1-a) region has 0.1579 and the pseudo exhaustive test patterns are least generated and the rate of test pattern is 1.22%, compared with exhaustive testing. (1-a) region has 0.2368 and this results exhibits the optimal performance of the singular cover and consistency operation. Applying the generated pseudo exhaustive test patterns to the stuck-at faults existing on the inputs and internal nodes in CUT, we verified this performance on the output. Thus, it is confirmed that functional testing of the proposed path sensitization algorithm is very useful.

• The KIPS Transactions:PartD
• /
• v.13D no.1 s.104
• /
• pp.11-22
• /
• 2006

As we are to deal with more and more XML documents, research on storing and managing XML documents in databases are actively conducted. Employing RDBMS or ORDBMS as a repository of XML documents is currently regarded as most practical. The query results out of XML documents stored in databases could be cached for query performance though the cost of cache consistency against the update of the underlying data is incurred. In this paper, we assume that an ORDBMS is used as a repository for the XML query cache as well as its underlying XML documents, and that XML query cache is refreshed in a deferred way with the update log. When the same XML document was updated multiple times, the deferred refresh of the XML query cache may Bet inefficient. We propose an algorithm that removes or filters such duplicate updates. Based on that, the optimal SQL statements that are to be executed for XML query cache consistency are generated. Through experiments, we show the efficiency of our proposed deferred refresh of XML query cache.

• Journal of Power Electronics
• /
• v.10 no.5
• /
• pp.548-554
• /
• 2010

In this paper, a 1.5 kW Interior Permanent Magnet Synchronous Generator (IPMSG) with a power conditioner for the grid integration of a variable-speed wind turbine is developed. The power-conditioning system consists of a series-type 12-pulse diode rectifier powered by a phase shifting transformer and then cascaded to a PWM voltage source inverter. The PWM inverter is utilized to supply sinusoidal currents to the utility line by controlling the active and reactive current components in the q-d rotating reference frame. While the q-axis active current of the PWM inverter is regulated to follow an optimized active current reference so as to track the maximum power of the wind turbine. The d-axis reactive current can be adjusted to control the reactive power and voltage. In order to track the maximum power of the wind turbine, the optimal active current reference is determined by using a simple MPPT algorithm which requires only three sensors. Moreover, the phase angle of the utility voltage is detected using a simple electronic circuit consisting of both a zero-crossing voltage detecting circuit and a counter circuit employed with a crystal oscillator. At the generator terminals, a passive filter is designed not only to decrease the harmonic voltages and currents observed at the terminals of the IPMSG but also to improve the generator efficiency. The laboratory results indicate that the losses in the IPMSG can be effectively reduced by setting a passive filter at the generator terminals.

• The KIPS Transactions:PartA
• /
• v.13A no.1 s.98
• /
• pp.79-86
• /
• 2006

Recently the tagSNP selection problem has been researched for reducing the cost of association studies between human's diversities and SNPs. General approach for this problem is that all of SNPs are separated into appropriate blocks and then tagSNPs are chosen in each block. Marsel in this paper is the system that involved the concept of linkage disequilibrium for overcoming the problem that the existing block partitioning approaches have short of biological meanings. In most approaches, the contiguous regions, which recombinations have LD coefficient |D'| and then tagSNP selection step is performed. And MarSel guarantees the minimum tagSNP selection using entropy-based optimal selection algorithm when tagSNPs are chosen in each block, and enables chromosome-level association studies using efficient memory management technique when input is very large-scale dataset that is impossible to be processed in the existing systems.

• Journal of KIISE:Databases
• /
• v.29 no.1
• /
• pp.15-26
• /
• 2002

Most work on multimedia storage systems has assumed that client will be serviced using a round-robin strategy. The server services the clients in rounds and each client is allocated a time slice within that round. Furthermore, most such algorithms are evaluated on the basis of a tightly coupled cost function. This is the basis of well-known algorithm such as FCFS, SCAN, SCAN-EDF, etc. In this paper, we describe a scheduling module called Request Unifier(RU) that takes as input, a set of client request, and a set of constraints on the desired performance such as client waiting time or maximum disk bandwidth, and a cost function. It produces as output a Unified Read Request(URR), telling the storage server which data items to read and when these data items to be delivered to the clients. Given a cost function, a URR is optimal if there is no other URR satisfying the constraints with a lower cost. We present three algorithms in this paper that can accomplish this kind of request merging and compare their performance through an experimental evaluation.

• The Journal of the Korea Contents Association
• /
• v.11 no.6
• /
• pp.51-58
• /
• 2011

One of the objectives of research on routing methods in wireless sensor networks is maximizing the energy life of sensor nodes that have limited energy. In this study, we tried to even energy use in a wireless sensor network by giving a weight to the transition probability of ACS(Ant Colony System), which is commonly used to find the optimal path, based on the amount of energy in a sensor and the distance of the sensor from the sink. The proposed method showed improvement by 46.80% on the average in energy utility in comparison with representative routing method GPSR (Greedy Perimeter Stateless Routing), and its residual energy after operation for a specific length of time was 6.7% more on the average than that in ACS.