• Annual Conference of KIPS
• /
• 2023.11a
• /
• pp.140-142
• /
• 2023

Federated learning (FL) is a new paradigm in machine learning (ML) that enables multiple devices to collaboratively train a shared ML model without sharing their local data. FL is well-suited for applications where data is sensitive or difficult to transmit in large volumes, or where collaborative learning is required. The Internet of Underwater Things (IoUT) is a network of underwater devices that collect and exchange data. This data can be used for a variety of applications, such as monitoring water quality, detecting marine life, and tracking underwater vehicles. However, the harsh underwater environment makes it difficult to collect and transmit data in large volumes. FL can address these challenges by enabling devices to train a shared ML model without having to transmit their data to a central server. This can help to protect the privacy of the data and improve the efficiency of training. In this view, this paper provides a brief overview of Fed-IoUT, highlighting its various applications, challenges, and opportunities.

• Journal of Forest and Environmental Science
• /
• v.39 no.4
• /
• pp.195-202
• /
• 2023

In this study, we sought to compare and evaluate the accuracy and predictive performance of machine learning algorithms for estimating the growth of individual Larix kaempferi trees in Gangwon Province, Korea. We employed linear regression, random forest, XGBoost, and LightGBM algorithms to predict tree growth using monitoring data organized based on different thinning intensities. Furthermore, we compared and evaluated the goodness-of-fit of these models using metrics such as the coefficient of determination (R²), mean absolute error (MAE), and root mean square error (RMSE). The results revealed that XGBoost provided the highest goodness-of-fit, with an R² value of 0.62 across all thinning intensities, while also yielding the lowest values for MAE and RMSE, thereby indicating the best model fit. When predicting the growth volume of individual trees after 3 years using the XGBoost model, the agreement was exceptionally high, reaching approximately 97% for all stand sites in accordance with the different thinning intensities. Notably, in non-thinned plots, the predicted volumes were approximately 2.1 m³ lower than the actual volumes; however, the agreement remained highly accurate at approximately 99.5%. These findings will contribute to the development of growth prediction models for individual trees using machine learning algorithms.

• Journal of Forest and Environmental Science
• /
• v.40 no.1
• /
• pp.35-42
• /
• 2024

The Japanese larch (Larix kaempferi [Lamb.] Carriere) is a major timber species in Korea. However, studies on bucking rates and merchantable logs of this species are insufficient in South Korea. To bridge these gaps, in this study, the bucking rate of Japanese larch (Larix kaempferi [Lamb.] Carriere) was computed and the number of long logs and merchantable log volumes were analyzed. Sample trees were bucked according to the log grade for trading, and collected from a forest field in Gangwon Province. The bucking rate of all Japanese larch logs was >89%. The highly profitable 2-4 logs of 3.6 m length from trees with ≤30 cm diameter at breast height (DBH) and 5-6 logs with ≥34 cm DBH were produced. The bucking rate of long logs was >84%; thus, Japanese larch was found to be suitable for the supply of high-grade timber. Additionally, to follow reasonable wood supply plans, merchantable volume tables were offered based on 3.6 m-long number of logs and small-end diameter classes. Understanding the proportion of merchantable log volumes, bucking rates, and the number of long and short logs has large-scale applications in practical forestry.

• Journal of Cardiovascular Imaging
• /
• v.30 no.3
• /
• pp.202-211
• /
• 2022

BACKGROUND: This study aims to investigate normal changes throughout aging of the heart in cardiac magnetic resonance (CMR) imaging in healthy volunteers. While type 2 diabetes mellitus is a frequent finding in the elderly population, also the influence of this circumstance in otherwise healthy persons is part of our study. METHODS: In this prospective single-center trial, 75 healthy subjects in distinct age groups and 10 otherwise healthy diabetics were enrolled. All subjects underwent functional, flow sensitive, native T2- and T1-mapping in a 1.5T CMR scanner. RESULTS: No differences in right and left ventricular ejection fractions were observed between aging healthy groups. Bi-ventricular volumes lowered significantly (p<0.001) between the age groups. There was also a significant decrease in myocardial T1 values, aortic distensibility, and left ventricular peak diastolic strain rates. There were no differences in T2 mapping and the other deformation parameters. Patients with type 2 diabetes mellitus had lower end-diastolic volume indexes; all the other measurements were comparable. CONCLUSIONS: Aging processes in the healthy heart involve a decrease in ventricular volumes, with ejection fractions remaining normal. Stiffening of the myocardium and aorta and a decrease in T1 values are potential indications of age-related remodeling. Type 2 diabetes mellitus seems to have no major influence on aging processes of the heart.

• The korean journal of orthodontics
• /
• v.54 no.5
• /
• pp.274-283
• /
• 2024

Objective: To determine the correlation between dentoskeletal parameters related to craniofacial morphology and the upper airway (UA) volume. Methods: Cone-beam computed tomography images of 106 randomly selected orthodontic patients were analyzed using NemoFab Ortho software. The dentoskeletal variables assessed were anterior facial height (AFH), posterior facial height (PFH), PFH/AFH ratio, hyoid position, maxillary width (MW), and palatal depth. The UA volume (evaluation in anatomical regions and as a whole) was also assessed using the same software. We also evaluated potential differences in UA variables between age and sex groups. The correlation between the dentoskeletal parameters and UA volume was calculated using the Pearson correlation coefficient (R). Analysis of variance and Student's t test were performed to assess differences between age and sex for UA variables. Statistical analyses were performed using SPSS software (version 26 for Windows). Results: This study found that PFH, AFH, and MW were the dentoskeletal parameters most strongly correlated with UA volume. However, the ANB angle did not show any significant correlation with UA volume. Additionally, differences in UA volumes were observed between age groups. Sex differences were found in both the "8-12" and "≥ 16" age groups for oropharyngeal and pharyngeal volumes. Conclusions: In conclusion, our findings indicate a significant correlation between UA volume and dentoskeletal parameters, particularly those related to facial height and MW.

• Journal of Korean Medical classics
• /
• v.21 no.4
• /
• pp.41-60
• /
• 2008

The acupuncture and moxibustion documentary characteristics of the "Dong-insuhyeolchimgudogyeong(銅人腧穴鍼灸圖經)" can be summarized into 5 parts such as the following. 1. The Index of Books lists the author of "Dong-indogyeong" as 'Wang-yuil(王惟一)' in some and 'Wang-yudeok(王惟德)' in others. Of these 'Wang-yudeok' is the name in the printed book and 'Yuil(惟一)' was the name it was changed into in order to make it different from the emperor Jinjong(眞宗) of the Song dynasty. Also, the sequence differs greatly among the original editions and sometimes even conflicts within itself, which makes very likely the possibility that the author of this book is not just one person, and that Wangyuil is just the main editor or screener. 2. The original printed edition of the Song(宋) dynasty has 3 volumes and the Cheonsoeng Seokgak(天聖 石刻) of the Song dynasty also has 3 volumes, but 'Hyeolsudosu(穴腧都數)', 'Sumyeongdanggyeolsik(修明堂訣式)', and 'Pichimgugyeol(避鍼灸訣)' were added to the 3rd volume. Of these three 'Pichimgugyeol(避鍼灸訣)' explains the 'Chimgupigijido(鍼灸避忌之圖)' of the 2nd volume in writing, 'Sumyeongdanggyeolsik(修明堂訣式)' explains the three human drawings from the front, side and lying down views in writing and was used to make bronze figures. 'Hyeolsudosu(穴腧都數)' was possibly used to draw acupuncture points when making bronze figures. During the Daejeong(大定) era of the Keum(金) dynasty Jinssi(陳氏) from the Pyeongsuseobang(平水書坊) revised this book and carved it into 5 volumes, and of those the 1st and 2nd volumes have characters that are very different from those of the rubbed copy of the stone carvings[石刻 拓本] from the Myeong(明) dynasty, which shows that it was taken from another book and not from the original text. The 3rd volume has added content from the stone carvings about accupuncture and moxibustion contraindication(鍼灸禁忌). During the Myeong dynasty the 3 volumes of the "Dong-indogyeong(銅人圖經)" were carved on wood based on the Jungtong Sukgan(正統 石刻), and there were many reprints with the contents that are almost all the same, but the order of the original edition were not all the same. When analyzed from many angles, the original printed copy was carved after the Jungtong(正統) of the Myung dynasty. 3. The content on transport points[輸血] in "Dong-indogyeong(銅人圖經)" comes mostly from books like "Oedae(外臺)" 'Myeongdang(明堂)', "Taepyeongseonghyebang(太平聖惠方)" 'Chimgyeong(鍼經)', "Yeongchugyeong(靈樞經)", Wangbing(王冰)'s annotation of "Somun(素問)", "Cheon-geumyobang(千金要方)" 'Chimgu(鍼灸)', and does not have much new content added by Wang-yuil(王惟一). However when Wang quotes past documents, little is summarized, some are edited and most do not have their origin indicated(the titles of other books quoted in the book are to give some supplementary explanation or are copied from another book). 4. There are many mistakes Wangyuil made in using documents and doing historical research and they can be seen in two fields. One is that he did not fix the errors of past documents, and the other is that he did not know the sequence of past documents or did not understand the meaning of the text. 5. The content Wangyuil newly added can mostly be seen in the few annotations and aneo(按語) under the original 'Geumbu(今附)'('Sinbu(新附)'). In order to see Wang's scholarly ideas and achievements we must rely on this part and the meridian entry of the transport points[輸血歸經] in the first part of the book. However we must not give credit to Wang for work that was done by former generations.

• The Korean Journal of Nuclear Medicine
• /
• v.37 no.2
• /
• pp.73-82
• /
• 2003

Objectives: A new software (Cardiac SPECT Analyzer: CSA) was developed for quantification of volumes and election fraction on gated myocardial SPECT. Volumes and ejection fraction by CSA were validated by comparing with those quantified by Quantitative Gated SPECT (QGS) software. Materials and Methods: Gated myocardial SPECT was peformed in 40 patients with ejection fraction from 15% to 85%. In 26 patients, gated myocardial SPECT was acquired again with the patients in situ. A cylinder model was used to eliminate noise semi-automatically and profile data was extracted using Gaussian fitting after smoothing. The boundary points of endo- and epicardium were found using an iterative learning algorithm. Enddiastolic (EDV) and endsystolic volumes (ESV) and election fraction (EF) were calculated. These values were compared with those calculated by QGS and the same gated SPECT data was repeatedly quantified by CSA and variation of the values on sequential measurements of the same patients on the repeated acquisition. Results: From the 40 patient data, EF, EDV and ESV by CSA were correlated with those by QGS with the correlation coefficients of 0.97, 0.92, 0.96. Two standard deviation (SD) of EF on Bland Altman plot was 10.1%. Repeated measurements of EF, EDV, and ESV by CSA were correlated with each other with the coefficients of 0.96, 0.99, and 0.99 for EF, EDV and ESV respectively. On repeated acquisition, reproducibility was also excellent with correlation coefficients of 0.89, 0.97, 0.98, and coefficient of variation of 8.2%, 5.4mL, 8.5mL and 2SD of 10.6%, 21.2mL, and 16.4mL on Bland Altman plot for EF, EDV and ESV. Conclusion: We developed the software of CSA for quantification of volumes and ejection fraction on gated myocardial SPECT. Volumes and ejection fraction quantified using this software was found valid for its correctness and precision.

• Progress in Medical Physics
• /
• v.21 no.2
• /
• pp.153-164
• /
• 2010

To determine the clinical target volumes considering vascularity and cellularity of tumors, the software was developed for mapping of the analyzed biological clinical target volumes on anatomical images using regional cerebral blood volume (rCBV) maps and apparent diffusion coefficient (ADC) maps. The program provides the functions for integrated registrations using mutual information, affine transform and non-rigid registration. The registration accuracy is evaluated by the calculation of the overlapped ratio of segmented bone regions and average distance difference of contours between reference and registered images. The performance of the developed software was tested using multimodal images of a patient who has the residual tumor of high grade gliomas. Registration accuracy of about 74% and average 2.3 mm distance difference were calculated by the evaluation method of bone segmentation and contour extraction. The registration accuracy can be improved as higher as 4% by the manual adjustment functions. Advanced MR images are analyzed using color maps for rCBV maps and quantitative calculation based on region of interest (ROI) for ADC maps. Then, multi-parameters on the same voxels are plotted on plane and constitute the multi-functional parametric maps of which x and y axis representing rCBV and ADC values. According to the distributions of functional parameters, tumor regions showing the higher vascularity and cellularity are categorized according to the criteria corresponding malignant gliomas. Determined volumes reflecting pathological and physiological characteristics of tumors are marked on anatomical images. By applying the multi-functional images, errors arising from using one type of image would be reduced and local regions representing higher probability as tumor cells would be determined for radiation treatment plan. Biological tumor characteristics can be expressed using image registration and multi-functional parametric maps in the developed software. The software can be considered to delineate clinical target volumes using advanced MR images with anatomical images.

• Journal of radiological science and technology
• /
• v.35 no.4
• /
• pp.327-333
• /
• 2012

Purpose: To analyze the correlation between dose volume histograms(DVH) based on organ outer wall contour and organ wall delineation for bladder and rectum, and to compare the doses to these organs with the absorbed doses at the bladder and rectum. Material and methods: Individual CT based brachytherapy treatment planning was performed in 13 patients with cervical cancer as part of a prospective comparative trial. The external contours and the organ walls were delineated for the bladder and rectum in order to compute the corresponding dose volume histograms. The minimum dose in 0.1 $cm^3$, 1 $cm^3$, 2 $cm^3$, 5 $cm^3$, 10 $cm^3$ volumes receiving the highest dose were compared with the absorbed dose at the rectum and bladder reference point. Results: The bladder and rectal doses derived from organ outer wall contour and computed for volumes of 2 $cm^3$, provided a good estimate for the doses computed for the organ wall contour only. This correspondence was no longer true when large volumes were considered. Conclusion: For clinical applications, when volumes smaller than 5 $cm^2$ are considered, the dose-volume histograms computed from external organ contours for the bladder and rectum can be used instead of dose -volume histograms computed for the organ walls only. External organ contours are indeed easier to obtain. The dose at the ICRU rectum reference point provides a good estimate of the rectal dose computed for volumes smaller than 2 $cm^2$ only for a midline position of the rectum. The ICRU bladder reference point provides a good estimate of the dose computed for the bladder wall only in cases of appropriate balloon position.

• Applied Biological Chemistry
• /
• v.28 no.2
• /
• pp.76-81
• /
• 1985

Miscible displacement techniques were used to investigate the influence of the organic matter treatment on the mobility of Cd, Zn, and Cu through soil columns. The heavy metals moved most readily through the Bonryang soil (Typic Udifluvents) of relatively low in CEC, pH, and organic matter content. Most parts of Cd and Zn eluted within 7 pore volumes, but Cu eluted between 5 and 15 pore volumes. Although the Gangseo soil (Aquatic Eutrochrepts) had lower in CEC and organic matter content than the Gyorae soil (Typic Distrandepts), the heavy metals moved faster through the Gyorae soil than through the Gangseo soil. Cu eluted more slowly and in smaller quantities than Cd and Zn from the Bonryang soil, but did not eluted from the Gangseo and the Gyorae soils at all during the experimental period. The motility of the heavy metals from the Bonryang and the Gangseo soils was in the order of Cd>Zn>Cu, but that of the Gyorae soil was in the order of Zn>Cd>Cu. Cd and Zn eluted after 5 and 20 pore volumes respectively, from the Bonryang soil treated with 3% compost but Cu did not elute even after 30 pore volumes were collected. By 7% compost treatment only small amountssof Cd eluted after 20 pore volumes. The liming of the Bonryang soil retarded the mobility of Cd, Zn, and Cu. Humic acid treatment did not reduce the motility of the Cd and Zn to the extent observed in the Bonryang soil with compost, but reduced a little motility of Cu.