• Progress in Medical Physics
• /
• v.8 no.1
• /
• pp.25-29
• /
• 1997

Ir-192 source activity for high dose rate brachytherapy is measured using Farmertype ionization chamber. The source-to-chamber distance is 10 cm and the measured charge unit is converted to activity unit. The measured values are compared to the values provided from vendor. Because of time dependency of Ir-192 source activity, the activities are regularly checked and compared to calculated values. As the accuracy of Ir-192 source activity is depend on the mechanical measurement setup, we estimated the precision of remote controlled source dwell position using home-made device and film scanner. The difference between measured and predicted dwell position is within 1 mm. As a result, the errors of source activity are 0.7${\pm}$1.5 ％ for measured and vendor-provided values and 0.l${\pm}$1.2％ for measured and time-dependent calculated vlaues. In conclusion, our measured activity has been comparable to the values provided from vendor and our brachytherapy unit has been very accurate until now. Regular quality control of brachytherapy is essential for successful treatment which depends on the accuracy of source position and activity.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1182-1182
• /
• 2001

In this study, the newly constructed optical measurement system, which was mainly composed of a parametric tunable laser and a near infrared photoelectric multiplier, was introduced to clarify the optical characteristics of wood as discontinuous body with anisotropic cellular structure from the viewpoint of the time-of-flight near infrared spectroscopy (TOF-NIRS). The combined effects of the cellular structure of wood sample, the wavelength of the laser beam λ, and the detection position of transmitted light on the time resolved profiles were investigated in detail. The variation of the attenuance of peak maxima At, the time delay of peak maxima Δt and the variation of full width at half maximum Δw were strongly dependent on the feature of cellular structure of a sample and the wavelength of the laser beam. The substantial optical path length became about 30 to 35 times as long as sample thickness except the absorption band of water. Δt ${\times}$ Δw representing the light scattering condition increased exponentially with the sample thickness or the distance between the irradiation point and the end of sample. Around the λ=900-950 nm, there may be considerable light scattering in the lumen of tracheid, which is multiple specular reflection and easy to propagate along the length of wood fiber. Such tendency was remarkable for soft wood with the aggregate of thin layers of cell walls. When we apply TOF-NIRS to the cellular structural materials like wood, it is very important to give attention to the difference in the light scattering within cell wall and the multiple specular-like reflections between cell walls. We tried to express the characteristics of the time resolved profile on the basis of the optical parameters for light propagation determined by the previous studies, which were absorption coefficient K and scattering coefficient S from Kubelka-Munk theory and n from nth power cosine model of radiant intensity. The wavelength dependency of the product of K/S and n, which expressed the light-absorbing and -scattering condition and the degree of anisotropy, respectively, was similar to that of the time delay of peak maxima Δt. The variation of the time resolved profile is governed by the combination of these parameters. So, we can easily find the set of parameters for light propagation synthetically from Δt.

• Geophysics and Geophysical Exploration
• /
• v.13 no.3
• /
• pp.227-234
• /
• 2010

The vertical response spectra using the observed ground motions from the recent more than 30 macro earthquakes were analysed and then were compared both to the seismic design response spectra (Reg Guide 1.60), applied to the domestic nuclear power plants, and to the Korean Standard Design Response Spectrum for general structures and buildings (1997). 176 vertical ground motions, without considering soil types, were used for normalization with respect to the peak acceleration value of each ground motion. The results showed that response spectrum had strong dependency on epicentral distance. The results also showed that the vertical response spectra revealed much higher values for frequency bands above 5~7 Hz than Reg. Guide (1.60). The results were also compared to the Korean Standard Response Spectrum for the 3 different soil types and showed that the vertical response spectra revealed much higher values for the frequency bands below 0.2 second (5 Hz) than the Korean Standard Response Spectrum (SD soil condition). These frequency-dependent spectral values could be related to the characteristics of the domestic crustal attenuation and the effect of each site amplification. However, through the qualitative improvements and quantitative enhancement of the observed ground motions, the conservation of vertical seismic design response spectrum should be considered more significantly for the frequency bands above 5 Hz.

• Applied Chemistry for Engineering
• /
• v.8 no.3
• /
• pp.395-402
• /
• 1997

The $SiO_2$ films were prepared by ECR(electron cyclotron resonance) plasma diffusion method, Deal-Grove model and Wolters-Zegers-van Duynhoven model were used to estimate the oxidation rate which was correlated with surface morphology for different orientation of Si(100) and Si(111). It was seen the $SiO_2$ thickness increased linearly with initial oxidation time. But oxidation rate slightly decrease with oxidation time. It was also shown that the oxidation process was controlled by the diffusion of the reactive species through the oxide layer rather than by the reaction rate at the oxide interface. The similar time dependency has been observed for thermal and plasma oxidation of silicon. From D-G model and W-Z model, the oxidation rate of Si(111) was 1.13 times greater than Si(100) because Si(111) had higher diffusion and reaction rate, these models more closely fits the experimental data. The $SiO_2$ surface roughness was found to be uniform at experimental conditions without etching although oxidation rate was increased, and to be nonuniform due to etching at experimental condition with higher microwave power and closer substrate distance.

• Journal of Broadcast Engineering
• /
• v.25 no.1
• /
• pp.1-12
• /
• 2020

In the medical field, numerical imbalance of data due to differences in disease prevalence is a common problem. It reduces the performance of a artificial intelligence network, leading to difficulties in learning a network with good performance. Recently, generative adversarial network (GAN) technology has been introduced as a way to address this problem, and its ability has been demonstrated by successful applications in various fields. However, it is still difficult to achieve good results in solving problems with performance degraded by numerical imbalances because the image resolution of the previous studies is not yet good enough and the structure in the image is modeled locally. In this paper, we propose a multi-scale conditional generative adversarial network based on attention mechanism, which can produce high resolution images to solve the numerical imbalance problem of chest X-ray image data. The network was able to produce images for various diseases by controlling condition variables with only one network. It's efficient and effective in that the network don't need to be learned independently for all disease classes and solves the problem of long distance dependency in image generation with self-attention mechanism.

• Progress in Medical Physics
• /
• v.13 no.1
• /
• pp.21-26
• /
• 2002

A practical calculation algorithm which calculates the relative output factor(ROF) for irregular shaped electron field has been developed and evaluated the accuracy of the algorithm. The algorithm adapted two-source model, which assumes that the electron dose can be express as sum of the primary source component and the scattered component from the shielding block. Original two-source model has been modified in order to make the algorithm simpler and to reduce the number of parameters needed in the calculation, while the calculation error remains within clinical tolerance range. The primary source is assumed to have Gaussian distribution, while the scattered component follows the inverse square law. Depth and angular dependency of the primary and the scattered are ignored ROF can be calculated with three parameters such as, the effective source distance, the variance of primary source, and the scattering power of the block. The coefficients are obtained from the square shaped-block measurements and the algorithm is confirmed from the rectangular or irregular shaped-fields used in the clinic. The results showed less than 1.0 ％ difference between the calculation and measurements for most cases. None of cases which have bigger than 2.1 ％ have been found. By improving the algorithm for the aperture region which shows the largest error, the algorithm could be practically used in the clinic, since one can acquire the 1011 parameter's with minimum measurements(5∼6 measurements per cones) and generates accurate results within the clinically acceptable range.

• Tunnel and Underground Space
• /
• v.20 no.6
• /
• pp.399-407
• /
• 2010

The horizontal response spectra using the observed ground motions from the recent more than 19 macro earthquakes were analysed and then were compared to both the seismic design response spectra (Reg Guide 1.60), applied to the domestic nuclear power plants, and the Korean Standard Design Response Spectrum for general structures and buildings (1997). 130 horizontal ground motions, without considering soil types, were used for normalization with respect to the peak acceleration value of each ground motion. The results showed that response spectrum have strong dependency on epicentral distance. The results also showed that the horizontal response spectra revealed much higher values for frequency bands above 5 Hz than Reg. Guide (1.60). The results were also compared to the Korean Standard Response Spectrum for the 3 different soil types and showed that the vertical response spectra revealed much higher values for the frequency bands below 0.3 second than the Korean Standard Response Spectrum (SD soil condition). These spectral values dependent on frequency could be related to characteristics of the domestic crustal attenuation and the effect of each site amplification. However, through the qualitative improvements and quantitative enhancement of the observed ground motions, the conservation of horizontal seismic design response spectrum should be considered more significantly for the frequency bands above 5 Hz.

• Journal of Intelligence and Information Systems
• /
• v.17 no.1
• /
• pp.111-125
• /
• 2011

Clustering is a process of grouping similar or relevant documents into a cluster and assigning a meaningful concept to the cluster. By this process, clustering facilitates fast and correct search for the relevant documents by narrowing down the range of searching only to the collection of documents belonging to related clusters. For effective clustering, techniques are required for identifying similar documents and grouping them into a cluster, and discovering a concept that is most relevant to the cluster. One of the problems often appearing in this context is the detection of a complex concept that overlaps with several simple concepts at the same hierarchical level. Previous clustering methods were unable to identify and represent a complex concept that belongs to several different clusters at the same level in the concept hierarchy, and also could not validate the semantic hierarchical relationship between a complex concept and each of simple concepts. In order to solve these problems, this paper proposes a new clustering method that identifies and represents complex concepts efficiently. We developed the Hierarchical Overlapping Clustering (HOC) algorithm that modified the traditional Agglomerative Hierarchical Clustering algorithm to allow overlapped clusters at the same level in the concept hierarchy. The HOC algorithm represents the clustering result not by a tree but by a lattice to detect complex concepts. We developed a system that employs the HOC algorithm to carry out the goal of complex concept detection. This system operates in three phases; 1) the preprocessing of documents, 2) the clustering using the HOC algorithm, and 3) the validation of semantic hierarchical relationships among the concepts in the lattice obtained as a result of clustering. The preprocessing phase represents the documents as x-y coordinate values in a 2-dimensional space by considering the weights of terms appearing in the documents. First, it goes through some refinement process by applying stopwords removal and stemming to extract index terms. Then, each index term is assigned a TF-IDF weight value and the x-y coordinate value for each document is determined by combining the TF-IDF values of the terms in it. The clustering phase uses the HOC algorithm in which the similarity between the documents is calculated by applying the Euclidean distance method. Initially, a cluster is generated for each document by grouping those documents that are closest to it. Then, the distance between any two clusters is measured, grouping the closest clusters as a new cluster. This process is repeated until the root cluster is generated. In the validation phase, the feature selection method is applied to validate the appropriateness of the cluster concepts built by the HOC algorithm to see if they have meaningful hierarchical relationships. Feature selection is a method of extracting key features from a document by identifying and assigning weight values to important and representative terms in the document. In order to correctly select key features, a method is needed to determine how each term contributes to the class of the document. Among several methods achieving this goal, this paper adopted the $x^2$�� statistics, which measures the dependency degree of a term t to a class c, and represents the relationship between t and c by a numerical value. To demonstrate the effectiveness of the HOC algorithm, a series of performance evaluation is carried out by using a well-known Reuter-21578 news collection. The result of performance evaluation showed that the HOC algorithm greatly contributes to detecting and producing complex concepts by generating the concept hierarchy in a lattice structure.

• Journal of the Korean Geographical Society
• /
• v.40 no.2 s.107
• /
• pp.221-241
• /
• 2005

Land Use and Land Cover Changes (LUCC) occur over a wide range of space and time scales, and involve complex natural, socio-economic, and institutional processes. Therefore, modelling and predicting LUCC demands an understanding of how various measured properties behave when considered at different scales. Understanding spatial and temporal variability of driving forces and constraints on LUCC is central to understanding the scaling issues. This paper aims to 1) assess the heterogeneity of land cover change processes over the landscape in northern Ghana, where intensification of agricultural activities has been the dominant land cover change process during the past 15 years, 2) characterise dominant land cover change mechanisms for various spatial scales, and 3) identify the optimal spatial scale for LUCC modelling in a savanna landscape. A multivariate statistical method was first applied to identify land cover change intensity (LCCI), using four time-sequenced NDVI images derived from LANDSAT scenes. Three proxy land use change predictors: distance from roads, distance from surface water bodies, and a terrain characterisation index, were regressed against the LCCI using a multi-scale hierarchical adaptive model to identify scale dependency and spatial heterogeneity of LUCC processes. High spatial associations between the LCCI and land use change predictors were mostly limited to moving windows smaller than 10$\times$10km. With increasing window size, LUCC processes within the window tend to be too diverse to establish clear trends, because changes in one part of the window are compensated elsewhere. This results in a reduced correlation between LCCI and land use change predictors at a coarser spatial extent. The spatial coverage of 5-l0km is incidentally equivalent to a village or community area in the study region. In order to reduce spatial variability of land use change processes for regional or national level LUCC modelling, we suggest that the village level is the optimal spatial investigation unit in this savanna landscape.

• Journal of the Korean Geographical Society
• /
• v.42 no.4
• /
• pp.488-505
• /
• 2007

The purposes of this research are twofold; 1) to verify spatial differences of tectonic movement using the spatial distribution of earthquakes, and 2) to infer mechanisms that generate spatial accumulation patterns of earthquakes in the Korean Peninsula. The first part of this sequential paper (Park, 2007) argues that the Korean Peninsula consists of four geostructural regions in which tectonic deformation and consequent geomorphological development patterns are different from each other Since this conclusion has been made by terrain analyses alone, it is necessary to verify this suggestion using other independent geophysical data. Because earthquakes are results of movement and deformation of land masses moving in different directions, the distribution of earthquake epicenters may be used to identify the direction and rates of land mass movement. This paper first analysed the spatial distribution of earthquakes using spatial statistics, and then results were compared with the spatial arrangement of geostructural regions. The spatial distribution of earthquakes in the Korean Peninsula can be summarized as the followings; firstly, the intensity of earthquakes shows only weak spatial dependency, and shows large difference even at adjacent regions. Secondly, the epicenter distribution has a clear spatial accumulation pattern, even though the intensity of earthquake shows a random pattern. Thirdly, the high density area of earthquakes shows a clear 'L' shape, passing through Pyeongannam-do, centered at Pyeongyang, and Hwanghae-do, Seosan and Pohang. The correlation coefficient between the density of earthquakes and distance from geostructral region boundaries is much higher than those between the density of fault lines and distance from tectonic division boundaries. Since fault lines and tectonic divisions in the Korean Peninsula are the results of long-term geological development, there is an apparent scale discrepancy to find significant correlations with earthquakes. This result verifies the research hypothesis that the Korean Peninsula is divided into four geostructral regions in which each has its own moving direction and spatial deformation characteristics. The existence of geostructural regions is also supported by the movement parrerns of land masses estimated from the GPS measurements. This conclusion is expected to provide a new perspective to understand the geomorphological developments and the earthquake occurrences in the Korean Peninsula.