• The KIPS Transactions:PartB
• /
• v.9B no.5
• /
• pp.571-578
• /
• 2002

In this paper, we propose a video sequence coding scheme called AMV (Auxiliary Motion Vector) to minimize error propagation caused by transmission errors over the Internet. Unlike the conventional coding schemes the AMY coder, for a macroblock in a frame, selects two best matching blocks among several preceding frames. The best matching block, called a primary block, is used for motion compensation of the destination macroblock. The other block, called an auxiliary block, replaces the primary block in case of its loss at the decoder. When a primary block is corrupted or lost during transmission, the decoder can efficiently and simply suppress error propagation to the subsequent frames by replacing the block with an auxiliary block. This scheme has an advantage of reducing both the number and the impact of error propagations. We implemented the proposed coder by modifying H.263 standard coding and evaluated the performance of our proposed scheme in the simulation. The simulation results show that AMV coder is more efficient than the H.263 baseline coder at the high packet loss rate.

• Smart Media Journal
• /
• v.1 no.3
• /
• pp.36-42
• /
• 2012

Recently, the smart applications, such as smart phone and smart TV, become a hot issue in IT consumer markets. In particular, the smart TV provides 3D video services, hence efficient coding methods for 3D video data are required. Three-dimensional (3D) video involves stereoscopic or multi-view images to provide depth experience through 3D display systems. Binocular cues are perceived by rendering proper viewpoint images obtained at slightly different view angles. Since the number of viewpoints of the multi-view video is limited, 3D display devices should generate arbitrary viewpoint images using available adjacent view images. In this paper, after we explain a view synthesis method briefly, we propose a new algorithm to compensate view synthesis errors around object boundaries. We describe a 3D warping technique exploiting the depth map for viewpoint shifting and a hole filling method using multi-view images. Then, we propose an algorithm to remove boundary noises that are generated due to mismatches of object edges in the color and depth images. The proposed method reduces annoying boundary noises near object edges by replacing erroneous textures with alternative textures from the other reference image. Using the proposed method, we can generate perceptually inproved images for 3D video systems.

• The Journal of Korean Society for Radiation Therapy
• /
• v.29 no.1
• /
• pp.19-26
• /
• 2017

Purpose: The application of density override is very important to minimize dose calculation errors by fiducial markers of metal material in proton treatment plan. However, density override with actual material of the fiducial marker could make problem such as inaccurate target contouring and compensator fabrication. Therefore, we perform density override with surrounding material instead of actual material and we intend to evaluate the usefulness of density override with surrounding material of the fiducial marker by analyzing the dose distribution according to the position, material of the fiducial marker and number of beams. Materials and Method: We supposed that the fiducial marker of gold, steel, titanium is located in 1.5, 2.5, 4.0, 6.0 cm from the proton beam's end of range using water phantom. Treatment plans were created by applying density override with the surrounding material and actual material of the fiducial marker. Also, a liver cancer patient who received proton therapy was selected. We located the fiducial marker of gold, steel, titanium in 0, 1.5, 3.5 cm from the proton beam's end of range and the treatment plans were created by same method with water phantom. Homogeneity Index(HI), Conformity Index(CI) and maximum dose of Organ At Risk(OAR) in Planning Target Volume(PTV) as the evaluation index were compared according to the material, position of the fiducial marker and number of beam. Results: The HI value was more decreased when density override with surrounding material of the fiducial marker was performed comparing with density override with actual material. Especially the HI value was increased when the fiducial marker was located farther from the proton beam's end of the range for a single beam and the fiducial marker's position was closer to isocenter for two or more beams. The CI value was close to 1 and OAR maximum dose was greatly reduced when density override with surrounding material of the fiducial marker was performed comparing with density override with actual material. Conclusion: Density override with surrounding material can be expected to achieve more precise proton therapy than density override with actual material of the fiducial marker and could increase the dose uniformity and target coverage and reduce the dose to surrounding normal tissues for the small fiducial markers used in clinical practice. Most of all, it is desirable to plan the treatment by avoiding the fiducial marker of metal material as much as possible. However, if the fiducial marker have on the beam path, density override of the surrounding material can be expected to achieve more precise proton therapy.

• Korean Journal of Ecology and Environment
• /
• v.55 no.4
• /
• pp.274-284
• /
• 2022

Through sample-size-based rarefaction analyses, we tried to suggest the appropriate degree of sample concentration and sub-sample extraction, as a way to estimate more accurate zooplankton species diversity when assessing biodiversity. When we collected zooplankton from three reservoirs with different environmental characteristics, the estimated species richness (S) and Shannon's H' values showed different changing patterns according to the amount of sub-sample extracted from the whole sample by reservoir. However, consequently, their zooplankton diversity indices were estimated the highest values when analyzed by extracting the largest amount of sub-sample. As a result of rarefaction analysis about sample coverage, in the case of deep eutrophic reservoir (Juam) with high zooplankton species and individual numbers, it was analyzed that 99.8% of the whole samples were represented by only 1 mL of sub-sample based on 100 mL of concentrated samples. On the other hand, in Soyang reservoir, which showed very small species and individual numbers, a relatively low representation at 97% when 10 mL of sub-sample was extracted from the same amount of concentrated sample. As such, the representation of sub-sample for the whole zooplankton sample varies depending on the individual density in the sample collected from the field. If the degree of concentration of samples and the amount of sub-sample extraction are adjusted according to the collected individual density, it is believed that errors that occur when comparing the number of species and diversity indices among different water bodies can be minimized.

• Journal of the Korean earth science society
• /
• v.24 no.3
• /
• pp.205-215
• /
• 2003

A precise and dense Bouguer anomaly is one of the most important data to improve the knowledge of our environment in the aspect of geophysics and physical geodesy. Besides the precise absolute gravity station net, we should consider two parts; one is to improve the precision in gravity measurement and correction of it, and the other is the density of measurement both in number and distribution. For the precise positioning, we have tested how we could use the GPS properly in gravity measurement, and deduced that the GPS measurement for 5 minutes would be effective when we used DGPS with two geodetic GPS receivers and the baseline was shorter than 40km. In this case we should use a precise geoid model such as PNU95. By applying this method, we are able to reduce the cost, time, and number of surveyors, furthermore we also get the benefit of improving in quality. Two kind of computer programs were developed to correct crossover errors and to calculate terrain effects more precisely. The repeated measurements on the same stations in gravity surveying are helpful not only to correct the drifts of spring but also to approach the results statistically by applying network adjustment. So we can find out the blunders of various causes easily and also able to estimate the quality of the measurements. The recent developments in computer technology, digital elevation data, and precise positioning also stimulate us to improve the Bouguer anomaly by more precise terrain correction. The gravity data of various sources, such as land gravity data (by Choi, NGI, etc.), marine gravity data (by NORI), Bouguer anomaly map of North Korea, Japanese gravity data, altimetry satellite data, and EGM96 geopotential model, were collected and processed to get a precise and dense Bouguer anomaly in and around the Korean Peninsula.

• Journal of Intelligence and Information Systems
• /
• v.16 no.4
• /
• pp.147-157
• /
• 2010

With the immergence of the iPhone, the interest in Smartphones is getting higher as services can be provided directly between service providers and consumers without the network operators. As the number of international tourists increase, individual tourists are also increasing. According to the WTO's (World Tourism Organization) prediction, the number of international tourists will be 1.56 billion in 2020,and the average growth rate will be 4.1% a year. Chinese tourists, in particular, are increasing rapidly and about 100 million will travel the world in 2020. In 2009, about 7.8 million foreign tourists visited Korea and the Ministry of Culture, Sports and Tourism is trying to attract 12 million foreign tourists in 2014. A research institute carried out a survey targeting foreign tourists and the survey results showed that they felt uncomfortable with communication (about 55.8%) and directional signs (about 21.4%) when they traveled in Korea. To solve this inconvenience for foreign tourists, multilingual servicesfor traffic signs, tour information, shopping information and so forth should be enhanced. The appearance of the Smartphone comes just in time to provide a new service to address these inconveniences. Smartphones are especially useful because every Smartphone has GPS (Global Positioning System) that can provide users' location to the system, making it possible to provide location-based services. For improvement of tourists' convenience, Seoul Metropolitan Government hasinitiated the u-tour service using Kiosks and Smartphones, and several Province Governments have started the u-tourpia project using RFID (Radio Frequency IDentification) and an exclusive device. Even though the u-tour or u-tourpia service used the Smartphone and RFID, the tourist should know the location of the Kiosks and have previous information. So, this service did not give the solution yet. In this paper, I developed a new convenient service which can provide location based information for the individual tourists using GPS, WiFi, and 3G. The service was tested at Insa-dong in Seoul, and the service can provide tour information around the tourist using a push service without user selection. This self-tour service is designed for providing a travel guide service for foreign travelers from the airport to their destination and information about tourist attractions. The system reduced information traffic by constraining receipt of information to tourist themes and locations within a 20m or 40m radius of the device. In this case, service providers can provide targeted, just-in-time services to special customers by sending desired information. For evaluating the implemented system, the contents of 40 gift shops and traditional restaurants in Insa-dong are stored in the CMS (Content Management System). The service program shows a map displaying the current location of the tourist and displays a circle which shows the range to get the tourist information. If there is information for the tourist within range, the information viewer is activated. If there is only a single resultto display, the information viewer pops up directly, and if there are several results, the viewer shows a list of the contents and the user can choose content manually. As aresult, the proposed system can provide location-based tourist information to tourists without previous knowledge of the area. Currently, the GPS has a margin of error (about 10~20m) and this leads the location and information errors. However, because our Government is planning to provide DGPS (Differential GPS) information by DMB (Digital Multimedia Broadcasting) this error will be reduced to within 1m.

• Journal of Sericultural and Entomological Science
• /
• v.8
• /
• pp.11-25
• /
• 1968

Various formulae for estimation of leaf production in mulberry trees were investigated and obtained. Four varieties of mulberry trees were used as the materials, and seven characters namely branch length. branch diameter, node number per branch, total branch weight, branch weight except leaves, leaf weight and leaf area, were studied. The formulae to estimate the leaf yield of mulberry trees are as follows: 1. Varietal differences were appeared in means, variances, standard devitations and standard errors of seven characters studied as shown in table 1. 2. Y$_1$=a$_1$X$_1$${\times}$P$_1$......(l) where Y$_1$ means yield per l0a by branch number and leaf weight determination. a$_1$.........leaf weight per branch. X$_1$.......branch number per plant. P$_1$........plant number per l0a. 3. Y$_2$=(a$_2$${\pm}$S. E.${\times}$X$_2$)+P$_1$.......(2) where Y$_2$ means leaf yield per l0a by branch length and leaf weight determination. a$_2$......leaf weight per meter of branch length. S. E. ......standard error. X$_2$....total branch length per plant. P$_1$........plant number per l0a as written above. 4. Y$_3$=(a$_3$${\pm}$S. E${\times}$X$_3$)${\times}$P$_1$.....(3) where Y$_3$ means of yield per l0a by branch diameter measurement. a$_3$.......leaf weight per 1cm of branch diameter. X$_3$......total branch diameter per plant. 5. Y$_4$=(a$_4$${\pm}$S. E.${\times}$X$_4$)P$_1$......(4) where Y$_4$ means leaf yield per 10a by node number determination. a$_4$.......leaf weight per node X$_4$.....total node number per plant. 6. Y$\sub$5/= {(a$\sub$5/${\pm}$S. E.${\times}$X$_2$)Kv}${\times}$P$_1$.......(5) where Y$\sub$5/ means leaf yield per l0a by branch length and leaf area measurement. a$\sub$5/......leaf area per 1 meter of branch length. K$\sub$v/......leaf weight per 100$\textrm{cm}^2$ of leaf area. 7. Y$\sub$6/={(X$_2$$\div$a$\sub$6/${\pm}$S. E.)}${\times}$K$\sub$v/${\times}$P$_1$......(6) where Y$\sub$6/ means leaf yield estimated by leaf area and branch length measurement. a$\sub$6/......branch length per l00$\textrm{cm}^2$ of leaf area. X$_2$, K$\sub$v/ and P$_1$ are written above. 8. Y$\sub$7/= {(a$\sub$7/${\pm}$S. E. ${\times}$X$_3$)}${\times}$K$\sub$v/${\times}$P$_1$.......(7) where Y$\sub$7/ means leaf yield estimates by branch diameter and leaf area measurement. a$\sub$7/......leaf area per lcm of branch diameter. X$_3$, K$\sub$v/ and P$_1$ are written above. 9. Y$\sub$8/= {(X$_3$$\div$a$\sub$8/${\pm}$S. E.)}${\times}$K$\sub$v/${\times}$P$_1$.......(8) where Y$\sub$8/ means leaf yield estimates by leaf area branch diameter. a$\sub$8/......branch diameter per l00$\textrm{cm}^2$ of leaf area. X$_3$, K$\sub$v/, P$_1$ are written above. 10. Y$\sub$9/= {(a$\sub$9/${\pm}$S. E.${\times}$X$_4$)${\times}$K$\sub$v/}${\times}$P$_1$......(9) where Y$\sub$7/ means leaf yield estimates by node number and leaf measurement. a$\sub$9/......leaf area per node of branch. X$_4$, K$\sub$v/, P$_1$ are written above. 11. Y$\sub$10/= {(X$_4$$\div$a$\sub$10/$\div$S. E.)${\times}$K$\sub$v/}${\times}$P$_1$.......(10) where Y$\sub$10/ means leaf yield estimates by leaf area and node number determination. a$\sub$10/.....node number per l00$\textrm{cm}^2$ of leaf area. X$_4$, K$\sub$v/, P$_1$ are written above. Among many estimation methods. estimation method by the branch is the better than the methods by the measurement of node number and branch diameter. Estimation method, by branch length and leaf area determination, by formulae (6), could be the best method to determine the leaf yield of mulberry trees without destroying the leaves and without weighting the leaves of mulberry trees.

• Journal of Intelligence and Information Systems
• /
• v.18 no.2
• /
• pp.29-45
• /
• 2012

Bond rating is regarded as an important event for measuring financial risk of companies and for determining the investment returns of investors. As a result, it has been a popular research topic for researchers to predict companies' credit ratings by applying statistical and machine learning techniques. The statistical techniques, including multiple regression, multiple discriminant analysis (MDA), logistic models (LOGIT), and probit analysis, have been traditionally used in bond rating. However, one major drawback is that it should be based on strict assumptions. Such strict assumptions include linearity, normality, independence among predictor variables and pre-existing functional forms relating the criterion variablesand the predictor variables. Those strict assumptions of traditional statistics have limited their application to the real world. Machine learning techniques also used in bond rating prediction models include decision trees (DT), neural networks (NN), and Support Vector Machine (SVM). Especially, SVM is recognized as a new and promising classification and regression analysis method. SVM learns a separating hyperplane that can maximize the margin between two categories. SVM is simple enough to be analyzed mathematical, and leads to high performance in practical applications. SVM implements the structuralrisk minimization principle and searches to minimize an upper bound of the generalization error. In addition, the solution of SVM may be a global optimum and thus, overfitting is unlikely to occur with SVM. In addition, SVM does not require too many data sample for training since it builds prediction models by only using some representative sample near the boundaries called support vectors. A number of experimental researches have indicated that SVM has been successfully applied in a variety of pattern recognition fields. However, there are three major drawbacks that can be potential causes for degrading SVM's performance. First, SVM is originally proposed for solving binary-class classification problems. Methods for combining SVMs for multi-class classification such as One-Against-One, One-Against-All have been proposed, but they do not improve the performance in multi-class classification problem as much as SVM for binary-class classification. Second, approximation algorithms (e.g. decomposition methods, sequential minimal optimization algorithm) could be used for effective multi-class computation to reduce computation time, but it could deteriorate classification performance. Third, the difficulty in multi-class prediction problems is in data imbalance problem that can occur when the number of instances in one class greatly outnumbers the number of instances in the other class. Such data sets often cause a default classifier to be built due to skewed boundary and thus the reduction in the classification accuracy of such a classifier. SVM ensemble learning is one of machine learning methods to cope with the above drawbacks. Ensemble learning is a method for improving the performance of classification and prediction algorithms. AdaBoost is one of the widely used ensemble learning techniques. It constructs a composite classifier by sequentially training classifiers while increasing weight on the misclassified observations through iterations. The observations that are incorrectly predicted by previous classifiers are chosen more often than examples that are correctly predicted. Thus Boosting attempts to produce new classifiers that are better able to predict examples for which the current ensemble's performance is poor. In this way, it can reinforce the training of the misclassified observations of the minority class. This paper proposes a multiclass Geometric Mean-based Boosting (MGM-Boost) to resolve multiclass prediction problem. Since MGM-Boost introduces the notion of geometric mean into AdaBoost, it can perform learning process considering the geometric mean-based accuracy and errors of multiclass. This study applies MGM-Boost to the real-world bond rating case for Korean companies to examine the feasibility of MGM-Boost. 10-fold cross validations for threetimes with different random seeds are performed in order to ensure that the comparison among three different classifiers does not happen by chance. For each of 10-fold cross validation, the entire data set is first partitioned into tenequal-sized sets, and then each set is in turn used as the test set while the classifier trains on the other nine sets. That is, cross-validated folds have been tested independently of each algorithm. Through these steps, we have obtained the results for classifiers on each of the 30 experiments. In the comparison of arithmetic mean-based prediction accuracy between individual classifiers, MGM-Boost (52.95%) shows higher prediction accuracy than both AdaBoost (51.69%) and SVM (49.47%). MGM-Boost (28.12%) also shows the higher prediction accuracy than AdaBoost (24.65%) and SVM (15.42%)in terms of geometric mean-based prediction accuracy. T-test is used to examine whether the performance of each classifiers for 30 folds is significantly different. The results indicate that performance of MGM-Boost is significantly different from AdaBoost and SVM classifiers at 1% level. These results mean that MGM-Boost can provide robust and stable solutions to multi-classproblems such as bond rating.

• The Journal of Korean Society for Radiation Therapy
• /
• v.34
• /
• pp.61-72
• /
• 2022

Purpose: Conventional CBCT(Cone-beam Computed-tomography) caused an error in the target volume due to organ movement in the area affected by respiratory movement. The purpose of this paper is to evaluate the usefulness of accuracy and time spent using the Gated CBCT function, which reduces errors when performing RGRT(respiratory gated radiation therapy), and to examine the appropriateness of phase. Materials and methods: To evaluate the usefulness of Gated CBCT, the QUASAR^TM respiratory motion phantom was used in the Truebeam STx^TM. Using lead marker inserts, Gated CBCT was scaned 5 times for every 20~80% phase, 30~70% phase, and 40~60% phase to measure the blurring length of the lead marker, and the distance the lead marker moves from the top phase to the end of the phase was measured 5 times. Using Cedar Solid Tumor Inserts, 4DCT was scanned for every phase, 20-80%, 30-70%, and 40-60%, and the target volume was contoured and the length was measured five times in the axial direction (S-I direction). Result: In Gated CBCT scaned using lead marker inserts, the axial moving distance of the lead marker on average was measured to be 4.46cm in the full phase, 3.11cm in the 20-80% phase, 1.94cm in the 30-70% phase, 0.90cm in the 40-60% phase. In Fluoroscopy, the axial moving distance of the lead marker on average was 4.38cm and the distance on average from the top phase to the beam off phase was 3.342cm in the 20-80% phase, 3.342cm in the 30-70% phase, and 0.84cm in the 40-60% phase. Comparing the results, the difference in the full phase was 0.08cm, the 20~80% phase was 0.23cm, the 30~70% phase was 0.10cm, and the 40~60% phase was 0.07cm. The axial lengths of ITV(Internal Target Volume) and PTV(Planning Target Volume) contoured by 4DCT taken using cedar solid tumor inserts were measured to be 6.40cm and 7.40cm in the full phase, 4.96cm and 5.96cm in the 20~80% phase, 4.42cm and 5.42cm in the 30~70% phase, and 2.95cm and 3.95cm in the 40~60% phase. In the Gated CBCT, the axial lengths on average was measured to be 6.35 cm in the full phase, 5.25 cm in the 20-80% phase, 4.04 cm in the 30-70% phase, and 3.08 cm in the 40-60% phase. Comparing the results, it was confirmed that the error was within ±8.5% of ITV Conclusion: Conventional CBCT had a problem that errors occurred due to organ movement in areas affected by respiratory movement, but through this study, obtained an image similar to the target volume of the setting phase using Gated CBCT and verified its usefulness. However, as the setting phase decreases, the scan time was increases. Therefore, considering the scan time and the error in setting phase, It is recommended to apply it to patients with respiratory coordinated stereotactic radiation therapy using a wide phase of 30-70% or more.

• The Korean Journal of Air & Space Law and Policy
• /
• v.31 no.2
• /
• pp.145-176
• /
• 2016

Aviation safety can be secured through regulations and policies of various areas and thorough execution of them on the field. Recently, for aviation safety management Korea is making efforts to prevent aviation accidents by taking various measures: such as selecting and promoting major strategic goals for each sector; establishing National Aviation Safety Program, including the Second Basic Plan for Aviation Policy; and improving aviation related legislations. Obstacle limitation surface is to be established and publicly notified to ensure safe take-off and landing as well as aviation safety during the circling of aircraft around airports. This study intends to review current aviation obstacle management system which was designed to make sure that buildings and structures do not exceed the height of obstacle limitation surface and identify its operating problems based on my field experience. Also, in this study, I would like to propose ways to improve the system in legal and regulatory aspects. Nowadays, due to the request of residents in the vicinity of airports, discussions and studies on aviational review are being actively carried out. Also, related ordinance and specific procedures will be established soon. However, in addition to this, I would like to propose the ways to improve shortcomings of current system caused by the lack of regulations and legislations for obstacle management. In order to execute obstacle limitation surface regulation, there has to be limits on constructing new buildings, causing real restriction for the residents living in the vicinity of airports on exercising their property rights. In this sense, it is regarded as a sensitive issue since a number of related civil complaints are filed and swift but accurate decision making is required. According to Aviation Act, currently airport operators are handling this task under the cooperation with local governments. Thus, administrative activities of local governments that have the authority to give permits for installation of buildings and structures are critically important. The law requires to carry out precise surveying of vast area and to report the outcome to the government every five years. However, there can be many problems, such as changes in the number of obstacles due to the error in the survey, or failure to apply for consultation with local governments on the exercise of construction permission. However, there is neither standards for allowable errors, preventive measures, nor penalty for the violation of appropriate procedures. As such, only follow-up measures can be taken. Nevertheless, once construction of a building is completed violating the obstacle limitation surface, practically it is difficult to take any measures, including the elimination of the building, because the owner of the building would have been following legal process for the construction by getting permit from the government. In order to address this problem, I believe penalty provision for the violation of Aviation Act needs to be added. Also, it is required to apply the same standards of allowable error stipulated in Building Act to precise surveying in the aviation field. Hence, I would like to propose the ways to improve current system in an effective manner.