• Journal of the Korean Society of Radiology
• /
• v.15 no.2
• /
• pp.153-158
• /
• 2021

X-ray image analysis is a very important field to improve the early diagnosis rate and prediction accuracy of periodontal disease. Research on the development and application of artificial intelligence-based algorithms to improve the quality of such dental X-ray images is being widely conducted worldwide. Thus, the aim of this study was to design a super-resolution algorithm for predicting periodontal disease and to evaluate its applicability in dental X-ray images. The super-resolution algorithm was constructed based on the convolution layer and ReLU, and an image obtained by up-sampling a low-resolution image by 2 times was used as an input data. Also, 1,500 dental X-ray data used for deep learning training were used. Quantitative evaluation of images used root mean square error and structural similarity, which are factors that can measure similarity through comparison of two images. In addition, the recently developed no-reference based natural image quality evaluator and blind/referenceless image spatial quality evaluator were additionally analyzed. According to the results, we confirmed that the average similarity and no-reference-based evaluation values were improved by 1.86 and 2.14 times, respectively, compared to the existing bicubic-based upsampling method when the proposed method was used. In conclusion, the super-resolution algorithm for predicting periodontal disease proved useful in dental X-ray images, and it is expected to be highly applicable in various fields in the future.

• Korean Journal of Environment and Ecology
• /
• v.22 no.4
• /
• pp.420-426
• /
• 2008

This study compared and analyzed EPT ratio, EPT richness and EPT number(number of species and individuals) of a total of 9 selected sites individually using EPT-group of aquatic insects in Musim stream, also compared similarity ratio between the sampling sites using UPGMA(Unweighted Paired Group Method using Arithmetic average) based on analysis subsequent to species structure of aquatic insects, and compared and analyzed a functional feeding groups(FFGs) of aquatic insects based on stream order. As a result, the aquatic insect group of 2,269 individuals is classified as 71 species, 36 families, and 8 orders, and EPT ratio and EPT richness at sites 3(0.72 and 3.89, respectively) were found to be relatively higher than those at the other sites; by contrast, those at site 9(0.03 and 0.09, respectively) were relatively low. EPT number of species and individuals was found to be lowest at site 9. As a result of UPGMA analysis, the sampling sites in Musim stream were divided into three groups, including A1(site 1, 7, 8), A2(site 2, 3, 4, 5, 6) and B(site 9). According to the results of FFGs, there was not shown specific difference in the number of functional feeding groups by each sites; however, there appeared a little more shredders and scrappers in the upper and mid stream comparing to other sampling sites while collector-filterers and collector-gatherers were found to be dominant in number in the lower stream. According to this research result, the sampling sites in Musim stream were found to have an effect on not only EPT richness, EPT ratio, and EPT number but also FFGs structure.

• The Korean Journal of Applied Statistics
• /
• v.33 no.4
• /
• pp.511-526
• /
• 2020

Single-cell RNA-sequencing (scRNA-seq) data consists of each cell's RNA expression extracted from large populations of cells. One main purpose of using scRNA-seq data is to identify inter-cellular heterogeneity. However, scRNA-seq data pose statistical challenges when applying traditional clustering methods because they have many missing values and high level of noise due to technical and sampling issues. In this paper, motivated by analyzing scRNA-seq data, we propose a novel spectral-based clustering method by imposing different weights on genes when computing a similarity between cells. Assigning weights on genes and clustering cells are performed simultaneously in the proposed clustering framework. We solve the proposed non-convex optimization using an iterative algorithm. Both real data application and simulation study suggest that the proposed clustering method better identifies underlying clusters compared with existing clustering methods.

• Journal of Korea Multimedia Society
• /
• v.12 no.1
• /
• pp.107-121
• /
• 2009

First reason for generation of standby power is because starting voltage must pass through from the source of electricity to IC. The second reason is due to current when IC is in operation. Purpose of this abstract is on structures of simple modules that automatically switch on or off through analysis of state on standby power and analysis of cutoff point patterns as well as application of algorithms. To achieve this, this paper is based on analysis of electric signals and modeling. Also, on/off cutoff criteria has been established for reduction of standby power. To find on/off cutoff point, that is executed algorithm of similar group and leading pattern group generation in the standby power state. Therefore, the algorithm was defined as an important parameter of the subtraction value of calculated between $1^{st}$ SCS, $2^{nd}$ SCS, and the median value of sampling coefficient per second from a wall outlet.

• Journal of Korea Multimedia Society
• /
• v.8 no.11
• /
• pp.1510-1519
• /
• 2005

We proposed a digital watermarking method that it is possible to identify the copyright because the watermark is detected in the first print-scan and to protect a forgery because the watermark is not detected in the second print-scan. The proposed algorithm uses LPM and DFT transform for the robustness to the distortion of pixel value and geometrical distortion. This methods could improve watermark detection performance and image quality by selecting maximum sampling radius in LPM transform. After analyzing the characteristics of print-scan process, we inserted the watermark in the experimentally selected frequency bands that survives robustly to the first print-scan and is not detected in the second print-scan, using the characteristic of relatively large distortion in high frequency bands of DFT As the experimental result, the original proof is possible because average similarity degree 5.13 is more than the critical value 4.0 in the first print-scan. And the detection of forgery image is also possible because average similarity degree 2.76 is less than the critical value 4.0 in the second print-scan.

• Korean Journal of Environment and Ecology
• /
• v.27 no.6
• /
• pp.765-776
• /
• 2013

This study was conducted to find out availability of a urban forest (Eudeungsan) as a substitute place of Mudeungsan Provincial Park (MPP) in overuse problem and to identify related factors to place substitutability. During June-August of 2012, 232 samples who have ever visited both of the two places (i.e., MPP and Euduengsan) in Gwang-ju city were selected by using convenient sampling and snowball sampling methods. Self-administered questionnaire survey was conducted, and the valid responses from 225 samples (97.0%) were used for data analysis. 24.2% of respondents had "considerably" or "very much" intention to use the substitute place, and 65.5% of respondents showed "a little" or "somewhat" 10.3% of respondents didn't have any intention to visit Eudeungsan as a substitute place of MPP. Generally, socio-demographic and visiting characteristics were not related with intention to visit Eudeungsan as a substitute place of MPP. It was only found that respondents who largely visit 'alone' to Eudeungsan tended to have higher intention to visit the substitute place than those who visit 'in group.' Three factors of reasons to visit Eudeungsan (i.e.,Challenge/self-development: ${\beta}=0.35$; Accessibility/familiarity: ${\beta}=0.27$; Adventure: ${\beta}=-0.19$) influenced intention to visit the substitute place. Relationship between reasons to visit and intention to visit the substitute place was found. No direct relationship was found between perceived similarity of the places and intention to visit substitute place, and similarity of recreation experiences showed perfect mediating effects between the two variables.

• Ocean and Polar Research
• /
• v.33 no.3
• /
• pp.235-253
• /
• 2011

The annual variation of mesozooplankton community in the Jangmok harbor of Jinhae Bay was studied in relation to environmental variables. Sampling was carried out weekly from January to December 2009. During the study periods, mesozooplankton community consisted of 44 taxa and the annual mean abundance was 8308 inds. $m^{-3}$. The maximum abundance was observed to be 50043 inds. $m^{-3}$ in August and the minimum in April with 1013 inds. $m^{-3}$. Of these, Penilia avirostris, cirripedia larvae, Evadne tergestina, Acartia omorii, Oikopleura s, Paracalanus parvus s. l., Eurytemora pacifica, Podon s, Oithona s, and Acartia steueri were observed as dominant species in Jangmok bay and they also contributed to 79% of total mesozooplankton. According to non-metric multidimensional scaling (nMDS) and cluster analysis based on the mesozooplankton community data from each season, the community was divided into three groups. The first group included appearence species in winter and spring season, which is mainly dominated the copepod such as A. omorii and E. pacifica. The second and third group was composed with observed species in summer and autumn, respectively. Based on the SIMPER (similarity percentages), P. avirostris in summer and cirripedia larvae in autumn were significantly dominated. Our results indicate that although the mesozooplankton abundances in Jangmok harbor fluctuated abruptly, its annual variation was strongly influenced by water temperature.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.7
• /
• pp.59-65
• /
• 2011

Korea DMB Service is popularized, and is used by many users. But latest display devices compared to the DMB content resolution support higher resolution and a variety of video resampling technologies has been used. Generally, subjective video quality is determined by object recognition rate in video, and increased as the edge space between objects are more clear. An edge is the boundary between an object and the background, and indicates the boundary between overlapping objects. the predicted direction in intra prediction used in H.264/AVC has the similarity up to 80% for the edge information. In the study, we propose an effective up-sampling mothed using the edge information that is extracted for the relationship between the intra prediction data and the DCT coefficient data of H.264 video encoding.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.106-106
• /
• 2015

The information of flow regimes continues to be norm in water resource and watershed management, in that stream flow regime is a crucial factor influencing water quality, geomorphology, and the community structure of stream biota. The objectives of this study were to estimate Korean stream flows from landscape variables, classify stream flow gages using hydraulic characteristics, and then apply these methods to ungaged biological monitoring sites for effective ecological assessment. Here I used a linear modeling approach (MLR, PCA, and PCR) to describe and predict seasonal flow statistics from landscape variables. MLR models were successfully built for a range of exceedance discharges and time frames (annual, January, May, July, and October), and these models explained a high degree of the observed variation with r squares ranging from 0.555 (Q95 in January) to 0.899 (Q05 in July). In validation testing, predicted and observed exceedance discharges were all significantly correlated (p<0.01) and for most models no significant difference was found between predicted and observed values (Paired samples T-test; p>0.05). I classified Korean stream flow regimes with respect to hydraulic and hydrologic regime into four categories: flashier and higher-powered (F-HP), flashier and lower-powered (F-LP), more stable and higher-powered (S-HP), and more stable and lower-powered (S-LP). These four categories of Korean streams were related to with the characteristics of environmental variables, such as catchment size, site slope, stream order, and land use patterns. I then applied the models at 684 ungaged biological sampling sites used in the National Aquatic Ecological Monitoring Program in order to classify them with respect to basic hydrologic characteristics and similarity to the government's array of hydrologic gauging stations. Flashier-lower powered sites appeared to be relatively over-represented and more stable-higher powered sites under-represented in the bioassessment data sets.

• Journal of Forest and Environmental Science
• /
• v.38 no.3
• /
• pp.141-151
• /
• 2022

Soil seed banks as natural storage of plant seeds play an important role in the maintenance and regeneration of watershed. Natural regeneration potential of the soil seed bank of Land use types (LUTs) in Ogun River watershed (ORW) was investigated. ORW was stratified using proportionate sampling technique into Guinea Savannah (GS), Rainforest (RF) and Swamp Forest (SF) Ecological Zones (EZs). Three LUTs: Natural Forest (NF), Disturbed Forest (DF) and Farmland (FL) were purposively selected in GS: GSNF, GSDF, GSFL; RF: RFNF, RFDF, RFFL and SF: SFNF, SFDF, SFFL, respectively. Systematic line transects was used in the laying of the sample plots. Sample plots of 25 m×25 m were established in alternate positions. Ten 1 m×1 m quadrats were randomly laid for soil core sampling from previously randomly selected ten plots. The core samples (10) were pooled per plot in each LUT and placed in individual trays. Ten trays with sterilized soil were used as control. The trays were watered regularly and checked for seedlings emergence fortnightly for 18 months. The experimental design used was 3×3 factorial experiments. ANOVA, Diversity index (H') and Similarity index (SI) were used to analyze the data. There was significant difference in seedling emergence among ecological zones and land use types (p<0.05). A total of 4,400 seedlings emerged from the soil samples. All species were distributed among 32 families. FL in the RF had the highest number of germinated seeds (705±37.33 seedlings) followed by DF in the RF (701±49.6 seedlings). The lowest emergence was in NF of the SF (199±28.41 seedlings). DF in the RF had highest number of species (34) distributed among 22 families. Emergence from soil seed bank of NF in ORW was generally with more of tree species than herbs that were predominant in FL and DF.