• Journal of Sensor Science and Technology
• /
• v.7 no.3
• /
• pp.188-196
• /
• 1998

The sensing and electrical characteristics of $WO_{3}$-based n-type semiconductor gas sensors are investigated. In normal air condition, $TiO_{2}$(4 wt. %)-doped $WO_{3}$-based sensor fabricated without any binder shows the grain boundary ( GB ) potential barrier height of 0.26 V. Sensors fabricated with alumina, PVA and silica sol binders show 0.17, 0.22 and 0.26 V of GB potential barrier height, respectively. In the ambience of 120 ppm $NO_{x}$ concentration, the GB potential barrier height of the sensor fablicated without binder is increased to 0.59 V. The sensors were fabricated with alumina, PVA, silica sol binders show 0.43, 0.66 and 0.52 V of potential barrier, respectively. Thus the variation of the potential barrier at GB is largest in the sensor fabricated with the PVA binder. This is found to be the main reason why the sensor fabricated with the PVA binder shows the best sensitivity. It is also found that the decrease of sensitivity at a temperature higher than the optimum operation temperature is due to the temperature dependence of the sensor resistance in normal air condition rather than the desorption of the adsorbed $NO_{x}$ gas particles. In the ambience of 250 ppm CO concentration, the GB potential barrier heights of the sensors fabricated without binder and with PVA binder are about 0.2 V showing negligible change compared to the case of normal air ambience. This fact indicates that these sensors are good candidates for the selective detection of $NO_{x}$ gas in the mixture of CO and $NO_{x}$ gases.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.3
• /
• pp.600-611
• /
• 1996

Recently, neural network-based speech recognition has been studied to utilize the adaptivity and learnability of neural network models. However, conventional neural network models have difficulty in the co-articulation processing and the boundary detection of similar phonmes of the Korean speech. Also, in case of using one phonotopic map, learning speed may dramatically increase and inaccuracies may be caused because homogeneous learning and recognition method should be applied for heterogenous data. Hence, in this paper, a neural net typewriter has been designed using a hierarchical self-organizing classifier(HSOC), and related algorithms are presented. This HSOC, during its learing stage, distributed phoneme data on hierarchically structured multiple phonotopic maps, using Kohonen's self-organizing feature maps(SOFM). Presented and experimented in this paper were the algorithms for deciding the number of maps, map sizes, the selection of phonemes and their placement per map, an approapriate learning and preprocessing method per map. If maps are divided according to a priorlinguistic knowledge, we would have difficulty in acquiring linguistic knowledge and how to alpply it(e.g., processing extended phonemes). Contrarily, our HSOC has an advantage that multiple phonotopic maps suitable for given input data are self-organizable. The resulting three korean phonotopic maps are optimally labelled and have their own optimal preprocessing schemes, and also confirm to the conventional linguistic knowledge.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.2 s.40
• /
• pp.25-33
• /
• 2006

We propose a method that automatically classifies the images into the object and non-object images. An object image is the image with object(s). An object in an image is defined as a set of regions that lie around center of the image and have significant color distribution against the other surround (or background) regions. We define four measures based on the characteristics of an object to classify the images. The center significance is calculated from the difference in color distribution between the center area and its surrounding region. Second measure is the variance of significantly correlated colors in the image plane. Significantly correlated colors are first defined as the colors of two adjacent pixels that appear more frequently around center of an image rather than at the background of the image. Third one is edge strength at the boundary of candidate for the object. By the way, it is computationally expensive to extract third value because central objects are extracted. So, we define fourth measure which is similar with third measure in characteristic. Fourth one can be calculated more fast but show less accuracy than third one. To classify the images we combine each measure by training the neural network and SYM. We compare classification accuracies of these two classifiers.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.4
• /
• pp.124-131
• /
• 2009

A reliable detection of regions in radiography is one of the most important task before the evaluation of defects on welded joints. The extracted features is to be classified into distinctive clusters for each segmented region. But conventional segmentation techniques give unsatisfactory results for this task due to the spatial superposition of intensity and low signal-to-ratio(SNR) in radiographic images. The usage of global or local processes not only provide the necessary noise resistance but also fail in classification of regions. In this paper, we presents an appropriate approach for segmentation of region-based indications in industrial Computed Radiography(CR) images. The geometric differences between welded and non-welded area which is generated on radiography as the representative regions(background, thickness, middle and welded region in steel tube image) have constructed the hierarchical structure. Although this structure is contaminated by noise, the scheme between regions can be selected by the help of local clustering based on distinctive geometric property of each region. Because of the geometric nature of the considered region and so that the region is selected layer by layer, and that the real class represents the boundary between regions, the vertical and horizontal clustering process in each layer must be judicious. In order to show the effectiveness of this approach, a comparative experiment of various segmentation method is performed on industrial steel tube CR images.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.43 no.7 s.349
• /
• pp.72-83
• /
• 2006

It is not a practical solution that the DDoS attacks or worm propagations are protected and responded within a domain itself because it clogs access of legitimate users to share communication lines beyond the boundary a domain. Especially, the DDoS attacks with spoofed source address or with bogus packets that the destination addresses are changed randomly but has the valid source address does not allow us to identify access of legitimate users. We propose a scheme of distributed network security management to protect access of legitimate users from the DDoS attacks exploiting randomly spoofed source IP addresses and sending the bogus packets. We assume that Internet is divided into multiple domains and there exists one or more domain security manager in each domain, which is responsible for identifying hosts within the domain. The domain security manager forwards information regarding identified suspicious attack flows to neighboring managers and then verifies the attack upon receiving return messages from the neighboring managers. Through the experiment on a test-bed, the proposed scheme was verified to be able to maintain high detection accuracy and to enhance the. normal packet survival rate.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.34 no.5
• /
• pp.503-514
• /
• 2016

Delineation of accurate object boundaries is crucial to provide reliable spatial information products such as digital topographic maps, building models, and spatial database. In LiDAR(Light Detection and Ranging) data, real boundaries of the buildings exist somewhere between outer-most points on the roofs and the closest points to the buildings among points on the ground. In most cases, areas of the building footprints represented by LiDAR points are smaller than actual size of the buildings because LiDAR points are located inside of the physical boundaries. Therefore, building boundaries determined by points on the roofs do not coincide with the actual footprints. This paper aims to estimate accurate boundaries that are close to the physical boundaries using airborne LiDAR data. The accurate boundaries are determined from the non-gridded original LiDAR data using initial boundaries extracted from the gridded data. The similar method implemented in this paper is also found in demarcation of the maritime boundary between two territories. The proposed method consists of determining initial boundaries with segmented LiDAR data, estimating accurate boundaries, and accuracy evaluation. In addition, extremely low density data was also utilized for verifying robustness of the method. Both simulation and real LiDAR data were used to demonstrate feasibility of the method. The results show that the proposed method is effective even though further refinement and improvement process could be required.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.20 no.3
• /
• pp.248-256
• /
• 2009

EB(Eigen-Beamforming) has widely been applied to MIMO(Multiple-Input Multiple-Output) systems to form beams which maximize the effective signal-to-interference plus noise ratio(SINR) of the receiver using the singular value decomposition(SVD) of the MIMO channel. However, the signal detection performance for the mobile station near the cell boundary is severely degraded and the transmission efficiency decreases due to the influence of the interference signal from the adjacent cells. In this paper, we propose an adaptive interference mitigation method for the EB transmission, and evaluate the reception performance. In particular, a reception strategy which adaptively utilizes optimal combining(OC) and minimum mean-squared error for Intercell spatial demultiplexing(MMSE-lSD) is proposed, and the reception performance is investigated in terms of the effective SINR and system capacity. For the average system capacity, the proposed adaptive reception demonstrates the performance enhancement compared to the conventional EB reception using the receiver beamforming vector, and up to 2 bps/Hz performance gain is achieved for mobile station located at the cell edge.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.3
• /
• pp.687-696
• /
• 2014

This paper presents a semantic object extraction method using user's stroke input, color, and depth information. It is supposed that a semantically meaningful object is surrounded with a few strokes from a user, and has similar depths all over the object. In the proposed method, deciding the region of interest (ROI) is based on the stroke input, and the semantically meaningful object is extracted by using color and depth information. Specifically, the proposed method consists of two steps. The first step is over-segmentation inside the ROI using color and depth information. The second step is semantically meaningful object extraction where over-segmented regions are classified into the object region and the background region according to the depth of each region. In the over-segmentation step, we propose a new marker extraction method where there are two propositions, i.e. an adaptive thresholding scheme to maximize the number of the segmented regions and an adaptive weighting scheme for color and depth components in computation of the morphological gradients that is required in the marker extraction. In the semantically meaningful object extraction, we classify over-segmented regions into the object region and the background region in order of the boundary regions to the inner regions, the average depth of each region being compared to the average depth of all regions classified into the object region. In experimental results, we demonstrate that the proposed method yields reasonable object extraction results.

• The KIPS Transactions:PartB
• /
• v.12B no.7 s.103
• /
• pp.811-822
• /
• 2005

Automated text categorization is to classify free text documents into predefined categories automatically and whose main goals is to reduce considerable manual process required to the task. The researches to improving the text categorization performance(efficiency) in recent years, focused on enhancing existing classification models and algorithms itself, but, whose range had been limited by feature based statistical methodology. In this paper, we propose RTPost system of different style from i.ny traditional method, which takes fault tolerant system approach and data mining strategy. The 2 important parts of RTPost system are reinforcement training and post-processing part. First, the main point of training method deals with the problem of defining category to be classified before selecting training sample documents. And post-processing method deals with the problem of assigning category, not performance of classification algorithms. In experiments, we applied our system to documents getting low classification accuracy which were laid on a decision boundary nearby. Through the experiments, we shows that our system has high accuracy and stability in actual conditions. It wholly did not depend on some variables which are important influence to classification power such as number of training documents, selection problem and performance of classification algorithms. In addition, we can expect self learning effect which decrease the training cost and increase the training power with employing active learning advantage.

• Tunnel and Underground Space
• /
• v.17 no.2 s.67
• /
• pp.128-138
• /
• 2007

At low in-situ stress, the continuity and distribution of natural fractures in rock mass predominantly control the failure processes. However at high in-situ stress, the failure process are affected and eventually dominated by stress-induced fractures preferentially growing parallel to the excavation boundary. This fracturing is often observed in brittle type of failure such as slabbing or spatting. Recent studies on the stress- or excavation-induced damage of rock revealed its importance especially in a highly stressed regime. In order to evaluate the brittle failure around a deep underground opening, physical model experiments were carried out. For the experiments a new tue triaxial testing system was made. According to visual observation and acoustic emission detection, brittle failure grades were classified under three categories. The test results indicate that where higher horizontal stress, acting perpendicular $(S_{H2})$ and parallel $(S_{H1})$ to the axis of the tunnel respectively, were applied, the failure grade at a constant vertical stress level (Sy) was lowered. The failure initiation stress was also increased with the increasing $S_{H1}\;and\;S_{H2}$. From the multi-variable regression on failure initiation stress and true triaxial stress conditions, $f(S_v,\;S_{H1},\;S_{H2})$ was proposed.