• Journal of IKEEE
• /
• v.12 no.1
• /
• pp.65-74
• /
• 2008

One of the popular methods used for pattern classification is the MBR (Memory-Based Reasoning) algorithm. Since it simply computes distances between a test pattern and training patterns or hyperplanes stored in memory, and then assigns the class of the nearest training pattern, it is notorious for memory usage and can't learn additional information from new data. In order to overcome this problem, we propose an incremental learning algorithm (iMPA). iMPA divides the entire pattern space into fixed number partitions, and generates representatives from each partition. Also, due to the fact that it can not learn additional information from new data, we present iMPA which can learn additional information from new data and not require access to the original data, used to train. Proposed methods have been successfully shown to exhibit comparable performance to k-NN with a lot less number of patterns and better result than EACH system which implements the NGE theory using benchmark data sets from UCI Machine Learning Repository.

• Journal of Broadcast Engineering
• /
• v.20 no.5
• /
• pp.778-781
• /
• 2015

The paper presents a new way of visual descriptor for deformable object retrieval on the basis of partition based description. The proposed descriptor technology partitions a given object into boundary area and interior area and extracts a descriptor from each area. The final descriptor combines these descriptors. From a given image, deformable object is segmented. The center position of the deformable object is calculated. The object is partitioned into N × N blocks on the basis of the given center position. Blocks are classified as boundary area and interior area depending on the pixels in the block. The proposed descriptor consists of extracted MPEG-7 dominant descriptors from both the boundary and interior area. The performance of proposed method is tested on a database of 1,973 handbag images constructed with view point changes. ARR (Average Retrieval Rate) is used for the retrieval accuracy of the proposed algorithm, compared with MPEG-7 dominant color descriptor.

• Journal of Broadcast Engineering
• /
• v.18 no.4
• /
• pp.643-646
• /
• 2013

We present an improved AdaBoost algorithm to avoid overfitting phenomenon. AdaBoost is widely known as one of the best solutions for object detection. However, AdaBoost tends to be overfitting when a training dataset has noisy samples. To avoid the overfitting phenomenon of AdaBoost, the proposed method divides positive samples into K clusters using k-means algorithm, and then uses only one cluster to minimize the training error at each iteration of weak learning. Through this, excessive partitions of samples are prevented. Also, noisy samples are excluded for the training of weak learners so that the overfitting phenomenon is effectively reduced. In our experiment, the proposed method shows better classification and generalization ability than conventional boosting algorithms with various real world datasets.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.40 no.5
• /
• pp.497-503
• /
• 2016

In this study, we formulate an acoustical topology optimization problem to optimally design a partition layout inside the expansion chamber of a muffler. The lower-limit insertion loss value at a target frequency is constrained, and the partition volume is selected as an object function. In this study, we calculate the insertion loss outside the duct, while to determine the noise-attenuation performance, we use the insertion loss value calculated inside the duct or transmission loss value obtained in a previous study. We employ the finite-element model for acoustical analysis, and we determine the transmission of an incident acoustic wave through each finite element using the functions of design variables that change continuously between "0" and "1." The rigid body elements, which totally reflect incident waves, build up partitions. Finally, we compare optimal topologies that depend on the target frequency and the allowed lower-limit value of insertion loss.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.10
• /
• pp.33-43
• /
• 2011

Recently, advances in mobile communication and location identifying technology of the moving object is evolving. Therefore, the location-based services based on request for service have increased and a variety of the indexing for the position management of moving objects has been studied. Because the index based on Euclidean space are no restriction of movement, it is difficult to apply to the real world. Also, there is additional cost to find adjacent road segments in road networks-based indexing. Existing studies of fixed static objects such as buildings or hospitals are not considered. In this paper, we propose an efficient road networks-based indexing for management of current positions. The proposed indexing partitions road networks by grids and has integrated road connection informations and manage separated extra indexing for fixed static objects. Through the experiment, we show that the proposed indexing based on road networks improves the performance of operation for search or update than existing indexing.

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 2004.09a
• /
• pp.29-32
• /
• 2004

Gaseous partitioning tracer test has been used for determining the volume and spatial distribution of residual non-aqueous phase liquid (NAPL) in the unsaturated soils. In this study, an experimental method for measuring the content of gas-exposed NAPL as well as that of total NAPL in a sand during air sparging was developed. Two different gaseous phase NAPL-partitioning tracers were used; n-pentane, with very low water solubility, was used as the tracer that partitions into NAPL that is only in contact with the mobile gas, and chloroform, with fairly good water solubility, was selected for detecting total NAPL content in the sand. Helium and difluromethanewere used as the non- reactive tracer and water-partitioning tracers, respectively. Using n-decane as a model NAPL (NAPL saturation of 0.018), 25.6% of total NAPL was detected by n-pentane at the water saturation of 0.68. Oniy 9.1% of total NAPL was detected by n-pentane at the water saturation of 0.84. This result implies that the quantity of gas-exposed NAPL increased about three times when the water saturation decreased from 0.84 to 0.68. At the water saturation of 0.68, more than 90% of total NAPL was detected by chloroform while 65.8% of total NAPL was detected by chloroform at the water saturation of 0.84. Considering that the removal rate of NAPL during air sparging for NAPL-contaminated aquifer is expected to be greatly dependent upon the spatial arrangement of NAPL phase with respect to the mobile gas, this new approach may provide useful information for investigating the mass transfer process during air-driven remedial processes fer NAPL-contaminated subsurface environment.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.37 no.7
• /
• pp.50-57
• /
• 2000

We propose a logic synthesis algorithm for the design of FPGAs operating at high speed. FPGA is a novel technology that provides programmability in the field. Because of short turnaround time and low manufacturing cost, FPGA has been noticed as an ideal device for system prototyping. Despite these merits, FPGA has drawbacks, namely low integration and long delay time comparing to ASIC. The proposed algorithm partitions a given circuit into subcircuits utilizing a kernel divisor such that the subcircuits can be performed at the same time, hence reducing the delay of the circuit. Experimental results on the MCNC benchmark show that the proposed algorithm is effective by generating circuits having 19.1% les delay on average, when compared to the FlowMap algorithm.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.46 no.5
• /
• pp.1-8
• /
• 2009

Since scan based testing is very efficient and widely used for testing large sequential circuits. However, since test patterns are serially injected through long scan chains, scan based testing requires very long test application time. Also, compared to the normal operations, scan shifting operations drastically increase power consumption. In order to solve these problems, this paper presents a new scan architecture for both test application time and test power reduction. The proposed scan architecture partitions scan chains into several segments and bypasses some segments which do not include any specified bit. Since bypassed segments are excluded from the scan shifting operation, the test application time and test power can be significantly reduced.

• The Journal of the Korea Contents Association
• /
• v.16 no.7
• /
• pp.667-680
• /
• 2016

As distributed computing platforms like Hadoop MapReduce have been developed, it is necessary to perform the conventional query processing techniques, which have been executed in a single computing machine, in distributed computing environments efficiently. Especially, studies on similarity join query processing in distributed computing environments have been done where similarity join means retrieving all data pairs with high similarity between given two data sets. But the existing similarity join query processing schemes for distributed computing environments have a problem of skewed computing load balance between clusters because they consider only the data transmission cost. In this paper, we propose Matrix-based Load-balancing Algorithm for efficient similarity join query processing in distributed computing environment. In order to uniform load balancing of clusters, the proposed algorithm estimates expected computing cost by using matrix and generates partitions based on the estimated cost. In addition, it can reduce computing loads by filtering out data which are not used in query processing in clusters. Finally, it is shown from our performance evaluation that the proposed algorithm is better on query processing performance than the existing one.

• Journal of Korea Multimedia Society
• /
• v.6 no.5
• /
• pp.769-777
• /
• 2003

We present a striping technique that MPEG-1 video streams ell a disk array can be efficiently played back at different resolution levels. For the MPEG-1 compression algorithm, the proposed multi-level encoding technique first partitions the parent video stream in the temporal dimension. Each frame in the sub-stream is then Partitioned in the chroma dimension yielding a low resolution and a residual component. The multimedia server stores blocks of different components on consecutive disks in a round robin manner. As a result, the lower the resolution level being maintained, the smaller is the number of disks accessed by each client. To effectively utilize a disk array and to maximize the number of clients that can be serviced simultaneously, the proposed technique interleaves the storage of the component of sub-streams among the disks in the array We empirically validate and evaluate this striping technique through simulation in order to show the improvement of its performance on the server.