Application of Magnetized Water System for Improvement of Reproductive Ability in Pigs

Lee, Sanghee Kangwon National University 2016 국내박사

A Study on the Relation between Type of Debt Financing and accounting conservatism

Sanghee Lee 고려대학교 대학원 2011 국내박사

This study examines the effects of firms' financing methods on accounting conservatism with a sample of 10,743 observations from listed firms over 10 years from 2000 to 2009. Specifically, the study divides debts into public and private debt, and further divides public debts into public bonds and private bonds, and attempts to test whether firms that rely more on public debt and public bonds are more conservative in accounting by recognizing economic losses in a more timely fashion when determining earnings. According to the efficient contract theory, accounting conservatism mitigates agency problems resulting from conflicts of interest between creditors and debtors. Public debt, usually held by many unspecified persons causes greater information asymmetry between creditors and debtor firms and thus incur more agency cost than private debts held by a small number of contracted parties who have greater information access and greater ability to control their debtor firms. Thus public󰠏debt creditors rely more on publicly available accounting information than do private󰠏debt creditors. Thus, it is expected that managers rely on conservative accounting to reduce agency costs and to meet the demand for information from creditors. This tendency is expected to be greater for firms with higher public󰠏debt ratios because of the higher agency costs, due to the greater information asymmetry and the fact that creditors rely more on publicly available financial statements. In this context, this study attempts test whether accounting conservatism shows a significant level of difference depending on the degree of reliance on different financing methods, such as public debt and private debt. The results are consistent regardless of the analytical model used. The findings of this study are as follows. First, firms with higher debt ratios are more conservative in their accounting, but not at a statistically significant level. To further test the association between accounting conservatism and debt type, this study divides debts by debt type and conducts tests to determine what effect each debt type has on accounting conservatism. Second, firms with higher public󰠏debt ratios are more conservative in their accounting practices. These results support the initial expectations of the study that conservatism prevails more in firms with higher public󰠏debt ratios that rely more on publicly available accounting information. Regarding public󰠏debt financing, firms with higher public󰠏bond ratios are more conservative in accounting. These results indicate that there are also differing degrees of conservatism between public bonds and private bonds, thus adding further confirmatory evidence on the role of conservatism in the efficient contract theory. This study makes the following contributions. First, by providing empirical evidence with regard to the association between debt financing methods and accounting conservatism, this study presents additional evidence on the role of conservatism in the efficient contract theory. Specifically, this study is meaningful in that it finds differing degrees of accounting conservatism even in the same market, depending on debt󰠏financing methods, as well as on country (Ball et al. 2008) or depending on whether the debtor firms are listed public companies (Ball and Shivakumar 2005). Second, the study findings have implications on the relationship between information demand on the part of creditors and corresponding patterns of financial reporting on the part of management. Nikolaev (2010) reports that there is a positive association between reliance on debt covenants in public󰠏debt contracts and accounting conservatism, and argues that the existence of private debts weakens the association. The findings of this study are consistent with Nikolaev (2010) and further show that an association exists between types of corporate bonds, namely, public and private bonds.

Molecular mechanism of human mismatch repair initiation

Lee, Sanghee UMI Dissertation Publishing ProQuest LLC 2015 국내박사

Development of aerosol retrieval algorithm over East Asia from TANSO-CAI measurements onboard GOSAT

Lee, Sang-hee Graduate School, Yonsei University 2018 국내박사

지속적인 온실가스 모니터링을 위한 최초의 위성이 2009년 1월 23일 일본에서 위성 GOSAT이 발사되었으며, GOSAT에는 이산화탄소 관측을 위한 TANSO-FTS와 구름과 에어러솔 정보 제공을 위한 TANSO-CAI 센서가 탑재되었다. 이 중, TANSO-CAI는 총 4개의 채널(한 개의 자외선 채널과 2개의 가시 영역 채널, 1개의 근적외 영역 채널)로 구성되어 있다. 본 연구논문에서는 TANSO-CAI의 관측 자료를 이용하여 에어러솔 광학 특성을 산출하는 알고리즘이 개발되었으며, 동아시아 영역에서의 에어러솔 광학 특성 정보를 산출한다. 개발된 알고리즘에는 에어러솔 유형 구분에 따른 정확한 에어러솔 광학 정보 산출을 위해 동아시아 산출 영역에 최적화된 에어러솔 모델을 채택하여 조견표를 작성하여 적용하였다. 이를 위해 에어러솔 모델 추출을 위해 사용된 관측 자료는 장기간 관측된 AERONET 태양광도계의 산출물이 이용되었다. 이와 더불어, 에어러솔 산출물의 정확도 향상을 위해 개발된 알고리즘에는 기존의 방법들과 새롭게 시도된 여러 방법들이 적용되었다. 첫째, 위성을 이용한 에어러솔 광학 특성 산출시 지표에 의한 산란 효과를 보정하기 위해 널리 사용되는 최소반사도법을 이용한 방법이 지표면 산출을 위해 적용되었다. 그러나, TANSO-CAI의 경우 3일에 한 번 동일 지역을 관측하는 한계와 관측 자료 내 줄무늬 현상으로 인한 오차로 인해 이전 30일 자료를 활용하여 산출하는 최소반사도법은 동일 경로의 관측자료를 이용하는 방법(산출 날짜 ± 11일)으로 수정된 방법이 적용되었다. 둘째, 구름, 해양 반짝임 및 탁수에 의한 오차 제거를 위한 경계값 방법이 적용되었다. TANSO-CAI의 경우 단 4개의 제한된 채널로 인해 구름 제거가 용이하지 않은 한계가 있다. 이를 극복하기 위해 각 채널들의 경계값 이용 및 식생지수(NDVI)를 활용한 방법이 적용되었다. 특히, 채널비(R(band2)/R(band3)와 NDVI를 이용하여 구름의 가장자리를 제거하는 새로운 방법이 제시되었으며, 이 방법은 해양에서 그 제거 효과가 탁월했다. 그러나, 이러한 효과에도 불구하고 이 방법은 식생변화가 활발한 육지에서는 적용하기 어려우며, 해양에서도 두꺼운 황사(severe dust)가 발생한 경우에는 구름과 유사한 신호로 인해 적용에 제한이 있다. 탁수 제거에서도 자외선 영역의 채널에서 산출된 지표반사도의 경계값 방법이 도입되었다. 셋째, 에어러솔 유형 구분을 위해 자외선 영역과 가시 영역의 두 채널의 차를 이용한 방법이 제시되었다. 일반적으로 위상 관측 자료를 이용한 에어러솔 유형 구분을 위해 자외선 영역의 두 채널을 이용한 에어러솔 인덱스(aerosol index, AI) 방법이 적용되지만, TANSO-CAI는 UV 채널을 1개만 가지고 있어 이 방법을 적용하는데 어려움이 따른다. 이를 극복하기 위해 본 연구논문에서는 제시된 방법은 에어러솔 유형에 따라 자외선 영역과 가시 영역에서 각 흡수도가 서로 달라 두 채널 간의 기울기를 이용하는 방법을 적용하였다. 이를 통해 산출된 에어러솔의 광학 두께는 다른 위성과의 비교 분석에서 안정적인 결과를 산출하였다. 그러나, 0.2 이하의 낮은 에어러솔 광학 두께에서는 다른 위성 산출물과 마찬가지로 에어러솔 유형 구분의 정확도가 다소 낮은 한계를 보였다. 넷째, TANSO-CAI의 복사휘도 보정을 위해 라디오메트릭 보정 계수(radiometric degradation factors, RDFs)가 적용되었다. TANSO-CAI의 경우 발사 이후 불규칙적이로 빠른 복사휘도의 신호감소가 나타났다. 정확한 에어러솔 광학 특성 산출의 위해 채널별 복사휘도 보정은 필수불가결한 요소이다. 이전 연구에서는 적용되지 않은 RDFs 적용을 통해 적용 전후의 값을 비교 제시하였으며, 산출 정확도 및 개선율을 제시하였다. 현재 일본 현업에서 TANSO-CAI의 관측 자료를 이용한 에어러솔 광학 특성 자료는 정확도가 낮아 TANSO-FTS의 이산화탄소 산출 알고리즘에 활용하지 않고 있다. 그러나, 본 연구에서 산출된 에어러솔 광학 특성은 연세대학교 이산화탄소 산출팀에서 개발한 YCAR 알고리즘의 입력 자료로 제공되어 이산화탄소 산출율 증대와 정확도 향상에 기여하였다. To provide improved aerosol properties for carbon dioxide (CO2) retrievals, an aerosol retrieval algorithm was developed from Thermal And Near infrared Sensor for carbon Observation-Cloud and Aerosol Imager (TANSO-CAI) launched in January 2009 onboard the Greenhouse gases Observing SATellite (GOSAT). CAI is a unique instrument to observe aerosol properties with ultraviolet channels in addition to essential channels of visible and near infrared. Despite of its unique advantage, aerosol information from TANSO-CAI has not been used operationally for GOSAT CO2 retrieval due to reasons including calibration issues. With the consideration of radiation degradation, this algorithm retrieves aerosol properties, e.g., aerosol optical depth (AOD), size information of aerosol, and aerosol type, in 0.1-degree grid resolution by look-up tables (LUTs). LUTs at each aerosol type provides optical properties of aerosol using inversion products from Aerosol Robotic NETwork (AERONET) sun-photometer observation over East Asia. To improve the accuracy of the TANSO-CAI aerosol algorithm, several new methods were introduced as follows: 1) a method using reflectance difference between ultraviolet (band 1) and visible (band 2) was suggested for selecting aerosol types from LUT depending on aerosol absorptivity; 2) a concept combined the normalized difference vegetation index (NDVI) and ratio of reflectance at band 2 to that at band 3 was introduced to remove clouds in TANSO-CAI reflectance; 3) surface reflectance was determined from the minimum composite method over 23-paths in the AOD retrievals; 4) a threshold test of estimated surface reflectance at band 2 was suggested to mask turbid water over ocean; and 5) seasonal and annual radiometric degradation factors (RDFs) was applied to correct the reflectance of TANSO-CAI. The TANSO-CAI aerosol algorithm provides aerosol properties such as AOD, size and type information for over four years from June 2009 to December 2013 in this study. Among them, the retrieved AODs were compared with those of AERONET and Aqua/MODIS level 2 datasets over land and ocean. Comparisons of AODs between AERONET and TANSO-CAI over East Asia show a good agreement with the correlation coefficient (R) 0.739 ± 0.046, root mean square error (RMSE) 0.232 ± 0.047, and linear regression slope 0.960 ± 0.083 for the entire period. Over ocean, the comparisons between Aqua/MODIS and TANSO-CAI for the same period show better agreement (R = 0.830 ± 0.047, RMSE = 0.140 ± 0.019, slope = 1.226 ± 0.063) than those of AERONET. However, the comparisons between Aqua/MODIS and TANSO-CAI over land have relatively lower correlation with approximately R = 0.67, RMSE = 0.40, slope = 0.77) than those over ocean. The retrieved aerosol optical properties in this study have been provided as input for increasing accuracy of CO2 retrieval with TANSO-Fourier Transform Spectrometer (FTS) measurements.

DahnGoon for orchestra

Lee, Sanghee UMI Dissertation Services 2006 국내박사

Prostatic Inflammation Triggers Voiding Dysfunction due to Neural Cross-Talk between the Prostate and Bladder

Lee, Sanghee The University of Wisconsin - Madison 2015 해외박사(DDOD)

Lower urinary tract symptoms (LUTS) and benign prostatic hyperplasia (BPH) are common in aging men. Prostatic inflammation is a very common histologic feature of the prostate in men with LUTS. To explain this association, inflammation has been postulated to induce hyperplastic growth. It has also been speculated that inflammation may sensitizes pelvic afferents and dorsal root ganglion (DRG) neurons innervating pelvic organs and thereby produce changes in voiding behavior. The purpose of my work was to test this speculation. We first developed an animal model of E.coli 1677-induced prostatic inflammation without cystitis by nitrofurantoin (NTF) administration. In histology and bacterial culture, it was evident that transurethral instillation of E.coli 1677 caused inflammatory response and infiltration in the prostate but not in the bladder. With our animal model, the effects of prostatic inflammation on voiding behavior were examined by metabolic cage test. We found that the inflamed group showed significantly increased voiding frequency at post-instillation day 1 (PID 1), PID 7 and PID 14 and decreased volume per void at all time points. Linked analysis of voiding frequency and volume per void revealed a preponderance of the low volume and high frequency (LVHF) voiding in mice with prostatic inflammation. In further studies, we demonstrated an anatomic basis for pelvic cross-talk involving the bladder and prostate by showing (1) shared afferent innervation of the prostate and bladder at spinal segments of T13, L1, L2, L6 and S1 and (2) convergent DRG neurons innervating both prostate and bladder in adult C57BL/6 mice. Taken together, our studies provide the first demonstration that prostatic inflammation induces a change in voiding behavior and provides neuroanatomical evidence to support a postulated mechanism of afferent sensitization and pelvic cross-talk for the association of prostate inflammation and LUTS in aging men. Our study demonstrates a neural cross-talk between the prostate and bladder and the role of prostatic inflammation on the voiding behavior in adult mouse. Our animal model enables researchers to determine a neuro-peptinergic mechanism critical for inflammation-mediated hyper-sensitization and its contribution on voiding behavior.

Clinical outcomes after apical surgery on the palatal root of the maxillary first molar using a palatal approach

이상희 Graduate School, Yonsei University 2020 국내박사

In the previous studies of apical surgery, few studies have discussed about the apical surgery on the palatal root of the maxillary first molar using a palatal approach. This study aimed to investigate the outcomes and complications associated with apical surgery on the palatal root of the maxillary first molar using a palatal approach. We searched for patients who underwent apical surgery with a palatal approach on the palatal root of the maxillary first molar between March 2010 and September 2017 by a single operator. A total of 46 teeth from 46 subjects were included, and they were examined at 6 and 12 months after surgery, and, annually, thereafter. Of the 46 treated cases, 11 were lost to follow-up. Thus, 35 cases were included in the analysis (recall rate 76%, average follow-up time 3.5 years). Success was considered achieved in 32 cases; 30 cases showed complete healing and two had incomplete healing. Three cases resulted in failure, all occurring within 1 year following surgery. To evaluate whether there was nerve damage on the surgical side, a pinprick test and cotton swab test were used with a 0-10 numerical rating scale (NRS), and the results were statistically compared using the Wilcoxon signed-rank test. The pinprick test results showed that all the NRS scores were higher than 7, and the responses on the surgical site were not statistically different from those on the contralateral site (P=0.109). All subjects showed normosensitivity to the cotton swab test. The success rate of apical surgery on the palatal root of the maxillary first molar using a palatal approach was 91.5%, which is similar to that observed in apical surgery performed on the other roots. Use of the palatal approach had an advantage in terms of accessibility and visibility, and the number of complications resulting from artery and nerve damage was minimal. 기존의 치근단 수술에 대한 연구에서는 상악 제1대구치 구개측 치근의 구개측 접근법에 관해서 거의 다루지 않고 있다. 이 연구는 구개측 접근법을 사용한 상악 제1대구치 구개측 치근의 치근단 수술에 대한 결과와 합병증 조사를 목표로 하였다. 2010년 3월부터 2017년 9월까지 한 명의 술자로부터 구개측 접근법을 사용한 상악 제1대구치 구개측 치근의 치근단 수술을 받은 환자들을 조사하였다. 총 46명의 환자의 46개 치아에 대하여 수술 후 6개월, 12개월 그리고 그 후에는 매년 조사를 시행하였다. 46건의 증례 중 11 증례 follow-up 소실로 35건의 증례로 분석 시행하였다.(리콜율 76%, 평균 follow-up 기간 3.5년) 중 32건을 성공으로 평가하였다; 30건은 complete healing으로 평가하였으며 나머지 2건은 incomplete healing로 평가하였다. 실패한 세 건의 경우 모두 수술 후 1년 내에 실패 양상을 나타내었다. 수술 부위의 신경손상을 평가하기 위해 0-10 numerical rating scale (NRS)를 통한 pinprick test와 cotton swab test를 시행하였으며 그 결과는 Wilcoxon signed-rank test를 사용하여 통계적으로 비교하였다. Pinprick test 결과 모든 환자에서 7 이상의 NRS 수치를 보였으며, 수술 부위와 반대측 부위의 반응이 통계적으로 차이를 보이지 않았다(P=0.109). 모든 환자가cotton swab test에서 normosensitivity를 보였다. 구개측 접근법을 사용한 상악 제1대구치의 구개측 치근의 치근단 수술의 성공률은 기존의 다른 치근에 대한 치근단 수술의 성공률과 유사한 91.5%를 보였다. 구개측 접근법은 치근에 대한 접근성과 시야확보 측면에서 이점이 있었으며 동맥 혹은 신경 손상으로 인한 합병증은 적었다.

Staggered DG method with deep learning-guided mesh refinement

이상희 Graduate School, Yonsei University 2022 국내석사

In this thesis, we present an expanded staggered discontinuous Galerkin method for the heterogeneous diffusion problems on general polygonal meshes. In groundwater hydrology, the diffusion coefficient may be close to zero; but it is barely possible to invert numerical zeros. An expanded formulation adopts gradient as a third explicit variable, in addition to the flux and pressure variables in the standard mixed method, to avoid the inversion of the diffusion tensor. We derive a priori and a posteriori error estimates. a priori error estimates show the optimal convergence rates in L2-norms for all of the three variables. Next, a residual-based a posteriori error estimator suitable for the method is suggested and proven to be reliable and efficient. The theoretical results are supported by three numerical experiments with a variety of meshes, ranging from structured to general meshes. We observe the optimal convergence behaviors, the robustness to mesh distortion including small edges and sliver elements, and the performance of the proposed a posteriori error estimator. Furthermore, the expanded SDG method is compared with the standard SDG method by defining and computing effectivity indices. When solving a problem with finite element methods (FEMs), meshes are often adaptively refined to obtain more accurate solutions. One of the most typical refinement strategies for polygons is mid-point (MP) refinement that connects the centroidal point to each mid-point of edges. Still, MP is likely to reproduce undesirable features of polygons such as small edges or hanging nodes. We deal with the problem by determining the ``shape" of a polygon not by exact information like coordinates but by rough visual data like images. The convolutional neural networks (CNNs) is used to resolve the issue, which is a well-known tool to identify and classify objects. Therefore, we construct a CNN model architecture for polygon classification and propose two different vertex selectors. To test the classifier, we conduct a numerical experiment with a singularity problem. The CNN-based approaches enable the elements to be ``reasonably good" in quality. We present the related theories and the numerical results throughout this thesis. 본 연구에서는 불균일 매질에서의 확산 방정식을 일반적인 다각형 요소망 위에서 효과적으로 해결하기 위한 확장된 스태거드 불연속 갤러킨 방법이 제시된다. 지하수 흐름 문제, 다공성 물질에서의 유동 문제 등에서 매질의 투과율이 낮은 경우에는 확산 계수가 0으로 수렴할 수 있다. 확산 계수의2 역수 계산이 필요한 기존의 혼합요소법으로는 이 문제를 정확하고 안정적으로 풀어낼 수 없게 된다. 확장된 혼합요소법에서는 표준 혼합요소법에서 사용하는 유동(flux) 변수와 압력(pressure) 변수 이외에 gradient 값을 보조 변수로서 도입하며 이러한 문제를 해결한다. 본 논문에서 확장된 스태거드 불연속 갤러킨 방법에 대한 방정식을 구성한 후에 선험적 오차 추정이 시행된다. 또한 수치적인 실험을 수행하여 L2-노름에 대한 최적수렴성을 이론적으로 뿐만 아니라 수치적으로도 확인한다. 더 나아가 이 방법에 적합한 잔차 기반 경험적 오차 추정량을 제안한다. 이 추정량의 신뢰도(reliability)와 효율성(efficiency)은 정리로서 증명된다. 적응적인 요소망 세분화에서 세분화 기준값으로 이 추정량을 쓸 수 있고, 세 가지 수치 예제에 대한 실험 결과로써 이 추정량과 적응적인 요소망 세분화의 효용성을 뒷받침한다. 여기서 규칙적인 요소망부터 임의로 형성된 다각 요소망까지 다양한 요소망을 사용하며 왜곡 요소망에 대한 견고성과 최적수렴의 경향성을 확인한다. 이뿐만 아니라 유효 지표를 도입하여 기존의 스태거드 불연속 갤러킨 방법과의 성능 비교 분석을 해본다. 한편, 가장 단순한 형태의 다각형 요소 세분화 전략은 무게 중심점을 각 변의 중점에 연결하는 것인데, 얇고 뾰족한 다각형이나 짧은 변 등을 계속해서 만들어낸다는 치명적인 단점을 갖고 있다. 요소망을 이루는 각 요소의 정칙성을 높이기 위하여 합성 곱 신경망(convolutional neural network, CNN)을 이용한 다각형 요소망 생성 모델을 구성하고, 특히 필요한 꼭짓점을 골라내는 두 가지 알고리즘을 제시한다. 이를 통해 불필요한 분할이나 분할을 줄이고 요소망의 정칙도를 개선한다. 마지막으로 수치적인 실험을 통하여 여러 알고리즘을 적용한 결과를 비교해보며 이론의 타당성을 입증한다.

Design and implementation of automatic person indexing and retrieval system based on the overlay text in the news interview video sequences

이상희 University of Ulsan 2019 국내박사

With the advent of the digital age, a vast amount of video data has been created by consumers and professionals over the last few decades. And the advances in the data capturing, storage, and communication technologies have made vast amounts of video data available to consumer and professional applications. The tremendous increase in the use of video data entails a need to develop effective methods to manage these multimedia resources by their content. In response to such demands, many researchers have been motivated to develop powerful indexing systems to ensure easy access to the relevant information, navigation, and organization in the vast repositories of video data. Recognizing the overlay text embedded in images and videos provides high-level semantic clues which enhance tremendously the automatic image and video indexing. These texts contain a more concise and direct description of the content of the video. Therefore, the overlay text plays an important role in the automated content analysis systems such as the scene understanding, indexing, browsing, and retrieval. Especially, the overlay text in the broadcasting news video sequences provides more meaningful of the content than any other type of videos. The detection and recognition of the overlay text have become a hot topic in news video analysis such as identification of person or place, name of the new-worthy event, date of the event, stock market, other news statistics, and news summaries. This dissertation proposes a novel approach to extract meaningful content information from the broadcasted news video sequences by collaborative integration of image understanding and natural language processing. As an actual example, we developed a person browser system that associates faces and overlaid name texts in videos. This is given news videos as a knowledge source, then automatically extracts face and name text association as content information. The proposed framework consists of the text detection module, the face detection module, and the person indexing module. For the preprocessing step, the proposed system makes the sub-clip based on the beginning frame for only focusing on the frames with overlay text. In the text detection module, the system executes overlay text detection and separates the name text line. And the system processes detection and extraction of the overlay text, and text recognition by optical character recognition (OCR). In the face detection module, the face thumbnail is extracted. The face detection module makes the representative thumbnail of the interviewee. And the person indexing module generates automatically the index metadata by named entity recognition (NER). And finally, the person indexing database is automatically made by combining the recognized text with the face thumbnail. The successful results of person information extraction reveal that the proposed methodology of integrated use of image understanding techniques and natural language processing technique is headed in the right direction to achieve our goal of accessing real contents of multimedia information.

(A) study on the relation between type of debt financing and accounting conservatism

이상희 Graduate School, Korea University 2011 국내박사