• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1032-1032
• /
• 2001

Artificial Neural Network (ANN) calibration techniques have been used commercially for agricultural applications since the mid-nineties. Global models, based on transmission data from 850 to 1050 nm, are used routinely to measure protein and moisture in wheat and barley and also moisture in triticale, rye, and oats. These models are currently used commercially in approx. 15 countries throughout the world. Results concerning earlier European ANN models are being published elsewhere. Some of the findings from that study will be discussed here. ANN models have also been developed for coarsely ground samples of compound feed and feed ingredients, again measured in transmission mode from 850 to 1050 nm. The performance of models for pig- and poultry feed will be discussed briefly. These models were developed from a very large data set (more than 20,000 records), and cover a very broad range of finished products. The prediction curves are linear over the entire range for protein, fat moisture, fibre, and starch (measured only on poultry feed), and accuracy is in line with the performance of smaller models based on Partial Least Squares (PLS). A simple bias adjustment is sufficient for calibration transfer across instruments. Recently, we have investigated the possible use of ANN for a different type of NIR spectrometer, based on reflectance data from 1100 to 2500 nm. In one study, based on data for protein, fat, and moisture measured on unground compound feed samples, dedicated ANN models for specific product classes (cattle feed, pig feed, broiler feed, and layers feed) gave moderately better Standard Errors of Prediction (SEP) compared to modified PLS (MPLS). However, if the four product classes were combined into one general calibration model, the performance of the ANN model deteriorated only slightly compared to the class-specific models, while the SEP values for the MPLS predictions doubled. Brix value in molasses is a measure of sugar content. Even with a huge dataset, PLS models were not sufficiently accurate for commercial use. In contrast an ANN model based on the same data improved the accuracy considerably and straightened out non-linearity in the prediction plot. The work of Mr. David Funk (GIPSA, U. S. Department of Agriculture) who has studied the influence of various types of spectral distortions on ANN- and PLS models, thereby providing comparative information on the robustness of these models towards instrument differences, will be discussed. This study was based on data from different classes of North American wheat measured in transmission from 850 to 1050 nm. The distortions studied included the effect of absorbance offset pathlength variation, presence of stray light bandwidth, and wavelength stretch and offset (either individually or combined). It was shown that a global ANN model was much less sensitive to most perturbations than class-specific GIPSA PLS calibrations. It is concluded that ANN models based on large data sets offer substantial advantages over PLS models with respect to accuracy, range of materials that can be handled by a single calibration, stability, transferability, and sensitivity to perturbations.

• Korean Journal of Remote Sensing
• /
• v.8 no.1
• /
• pp.27-43
• /
• 1992

The purpose of this study is to detect and evaluate the change of urban vegetation of Seoul during 1980s. Large areas covered with agricultural crops or forests were converted to residential and commercial areas, roads, schools, sports complexes, etc. There were also widespreas concerns on the deterioration of the quality of urban vegetation due to severe air pollution, overcrowding of nature parks, and idling of farm lands by land speculators. The image used for this study were MSS(Oct. 4, 1979) and TM(Apr. 26, 1990). The Green Vegetation Index of Kauth & Thomas(1976) was for the analysis. The GVI were resampled with 75$\times$75m grids and overlaid with the jurisdictional boundaries of 22 districts of Seoul. The results were reclassified to 6 classes, class 6 representing grids with the most vigorous vegetation or the best vegetation improvement in 1980s. The finding of this study can be summarized as follows : First, the most vigorous vigorous vegetation, in terms of GVI, of the 1979 image can be found at paddy fields located on alluvial near Han River. Broad-leaf forests located on hilly terrains have higher GVI than conifers located on the upper-parts of mountains. The average GVI of the northern part and southern part of Han River are 3.56 and 3.74, respectively. The main reason why the southern part has higher GVI is that there are more prime agricultural lands. Districts of Kangseo, Yangcheon, and Songpa have the highest percentage of grids of GVI class 6, and the percentages are 3.55 %, 3.47 %, and 2.69 %, respectively. Second, the most vigorous vegetation of the 1990 image can be found at the grass lands of the Yongsan golf club and the Sungsu horse racing track. The GVI of farm lands is lower than forest because most agricultural crops are at the early stage of growing season when the TM image was taken. The size of built-up area is much larger than of 1979. On the other hand, vegetation patches surrounded by developed area become smaller and have stronger contrast to surrounding area. The average GVI of the northern part and southern part of Han River are 3.57 and 3.51, respectively. The main reason why the southern part has lower GVI is the at more large-scale urban development projects were carried out in there during 1980s. Districts of Tobong, Nowon, and Seocho have the highest percentage of class 6, and the perecentages are 16.58 %, 10.14 %, and 8.50% respectively. Third, the change of urban vegetation in Seoul during 1980s are significant. Grids of GVI change classes 1 and 2, which represent severe vegetation loss, occupy 15.97% of Seoul. Three districts which lost the most vegetation are Yangcheon, Kangseo, and Songpa, where the percentages of GVI class 1 are 13.42%, 13.39% and 9.06%, respectively. The worst deterioration was mainly caused by residential developments. On the other hand, the vegetation of some part of Seoul improved in this period. Grids of GVI change classes 5 and 6 occupy 9.83 % of Seoul. Distircts of Jung, Yongsan, and Kangnam have the highest percentage of grids with GVI change classes 5 and 6, and their percentages are 22.31%, 19.17%, and 13.66%, respectively. The improvement of vegetation occurred in two areas. Forest vegetation is generally improving despite of concerns based on air pollution and heavy use by recreationists. Vegetation in open spaces established in riverside parks, large residential areas, and major public facilities are also improving.

• Journal of the Korean Society of Radiology
• /
• v.17 no.6
• /
• pp.965-975
• /
• 2023

Education in radiology involves a large portion of practical training, and it is difficult to conduct it through non-face-to-face online education. In this study, we utilized the Metaverse platform, which can replace hands-on education, to implement a practice room that is difficult to access through non-face-to-face classes into a virtual world, and then evaluated the satisfaction of learners after the practice class, the practicality of the Metaverse platform class, and the future orientation of the curriculum. I wanted to find out. Using the metaverse platforms ZEPETO (Build It) and ZEP (ZEP), the S University radiology department lab was implemented into a virtual world and used for students' classes. A total of 50 students were surveyed twice, divided into pre- and post-surveys, and all questions used a 5-point Likert scale. As a result of the study, satisfaction was low at 2.32 for education without using Metaverse virtual space, while education using virtual space was very high at 4.16. As a result of the analysis, the satisfaction level of the new education system and the practicality of the Metaverse platform classes are very high, and it is believed that it will be a more effective education platform when conducting additional education such as app explanations in the future.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.2 no.1
• /
• pp.109-131
• /
• 2007

The purpose of this study were to a) clarify the current practice of internal marketing by contracted food service management companies, b) analyze the effectiveness of current internal marketing practices. The survey was conducted between October 12 and November 2, 2002 with the 10 contracted food service management companies; 6 large, and 4 small & medium sized companies. No significant difference was existed with the categories of internal marketing. However significant differences were existed with the service education, sanitary education, periodic interviews, rewards for best employees, incentive payments, paid -leave allowances, job function allowances, and position classes including unit manager, dietician and cook of MBO. The actual administration of internal marketing strategies showed that, the significant differences were existed with education, paid-leave, and employee welfare systems among position classes. Also the significant differences were existed with all categories of internal marketing between large and small & medium sized companies. In general, small & medium sized companies score of the internal marketing performance was lower than that of large sized companies. Therefore, the segmented internal marketing strategies depending on the analysis not only demand for employees but also the size of company, were required to be an effectively and efficiently managed contracted food management companies.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.35 no.3
• /
• pp.187-194
• /
• 2017

The present study introduces a method to parallelize k-means clustering algorithm for fast unsupervised classification of large satellite imagery. Known as a representative algorithm for unsupervised classification, k-means clustering is usually applied to a preprocessing step before supervised classification, but can show the evident advantages of parallel processing due to its high computational intensity and less human intervention. Parallel processing codes are developed by using multi-threading based on OpenMP. In experiments, a PC of 8 multi-core integrated CPU is involved. A 7 band and 30m resolution image from LANDSAT 8 OLI and a 8 band and 10m resolution image from Sentinel-2A are tested. Parallel processing has shown 6 time faster speed than sequential processing when using 10 classes. To check the consistency of parallel and sequential processing, centers, numbers of classified pixels of classes, classified images are mutually compared, resulting in the same results. The present study is meaningful because it has proved that performance of large satellite processing can be significantly improved by using parallel processing. And it is also revealed that it easy to implement parallel processing by using multi-threading based on OpenMP but it should be carefully designed to control the occurrence of false sharing.

• Journal of Environmental Impact Assessment
• /
• v.8 no.3
• /
• pp.1-11
• /
• 1999

This study was carried out to evaluate urban environmental structure in Taegu metropolitan region(TMR) with factor analysis, fuzzy set theory, geostatistic and geographic information system(GIS). The factor analysis could choose the representative one out of multiple variables and simplify the evaluation of the urban environmental structure. The fuzzy approach is an attempt to model an aspect of human thinking previously neglected; it starts from the premise that humans don't represent classes of objects as fully disjoint but rather as sets where transitions from membership to non-membership is gradual. The Geographic Information System(GIS) could connect attributes of factor scores derived from factor analysis to digital map by a method so called 'Spatial join'. The results obtained were as follows: Urbanization appearance was concentrated in the large cities, and this appearance was partial extremely, therefore, there has been a structural gap between urban area and agricultural area which was unified into the urban area. All inclinations didn't become worse after sudden urbanization. For example, suburban agriculture was developed as a large scale in the region near the large cities. Then it encouraged farmers in changing their old cultivating methods to the latest ones. But many districts in urban fringe had symptom of urbanization, the districts which were located between large cities have been developed gradually because of urban sprawl, and played a role in connecting each city. Therefore, due to the urbanization, forestry area and agricultural land, well conserved in the agricultural town, could be easily destroyed. In a different way with the urbanization of the Seoul metropolitan region, that of TMR was centralized upon the center of each city, and it was also very partial. But, because so many regions have the potentiality of urbanization, hereafter, the urbanization process in each region is likely to be different multifariously according to the urban management methods.

• Animal cells and systems
• /
• v.1 no.2
• /
• pp.341-344
• /
• 1997

Receptor-mediated endocytosis has been suggested for a stage-specific transport mechanism of vitellogenin into the oocytes of a sabellid poly chaete, Pseudopotamilla occelata. Membrane proteins of oocytes of three size classes, including small (30-70 $\mu\textrm{m}$ in diameter), intermediate (70-140 $\mu\textrm{m}$ in diameter) and large (180-200 $\mu\textrm{m}$ in diameter), showed a atage-specific variation. Coelomic fluid proteins (CP), ass$\mu\textrm{m}$ed to be vitellogenin, consists of several proteins, which showed quite a different pattern from that of yolk proteins. Incorporation of $^{125}I$-CP into the oocytes of the intermediate size class almost linearly increases with time, showing a contrast to the pattern of the large size class, in which the incorporation is low and approaches a plateau, suggesting the vitellogenin transports by a regulated process only in the intermediate size class. Vitellogenin receptor proteins were identified to be 60 kDa and 68 kDa only in the intermediate size class by a ligand blotting test.

• Journal of KIISE:Software and Applications
• /
• v.27 no.5
• /
• pp.563-574
• /
• 2000

Recognition of numeral fields is a very important task for many document automation applications. Conventional methods are based on the two-steps process, segmentation of touching numerals and recognition of the individual numerals. However, due to a large variation of touching types this approach has not produced a robust result. In this paper, we present a new segmentation-free method for recognizing the two touching numerals. In this approach, two touching numerals are regarded as a single pattern coming from 100 classes ('00', '01', '02', ..., '98', '99'). For the test set, we manually extract two touching numerals from the data set of NIST numeral fields. Due to the limitation of conventional neural network in case of large-set classification, we use a modular neural network and Drove its superiority through recognition experimen.

• Genomics & Informatics
• /
• v.12 no.3
• /
• pp.87-97
• /
• 2014

Although the number of protein-coding genes is not highly variable between plant taxa, the DNA content in their genomes is highly variable, by as much as 2,056-fold from a 1C amount of 0.0648 pg to 132.5 pg. The mean 1C-value in plants is 2.4 pg, and genome size expansion/contraction is lineage-specific in plant taxonomy. Transposable element fractions in plant genomes are also variable, as low as ~3% in small genomes and as high as ~85% in large genomes, indicating that genome size is a linear function of transposable element content. Of the 2 classes of transposable elements, the dynamics of class 1 long terminal repeat (LTR) retrotransposons is a major contributor to the 1C value differences among plants. The activity of LTR retrotransposons is under the control of epigenetic suppressing mechanisms. Also, genome-purging mechanisms have been adopted to counter-balance the genome size amplification. With a wealth of information on whole-genome sequences in plant genomes, it was revealed that several genome-purging mechanisms have been employed, depending on plant taxa. Two genera, Lilium and Fritillaria, are known to have large genomes in angiosperms. There were twice times of concerted genome size evolutions in the family Liliaceae during the divergence of the current genera in Liliaceae. In addition to the LTR retrotransposons, non-LTR retrotransposons and satellite DNAs contributed to the huge genomes in the two genera by possible failure of genome counter-balancing mechanisms.

• The Journal of Korean Medicine
• /
• v.28 no.3 s.71
• /
• pp.232-243
• /
• 2007

Objective : This study researched the application of Sa-am acupuncture for the treatment of hiccup. Methods : We investigated the literature for Sa-am acupuncture treatment and traditional acupuncture treatment for hiccup. Result & Conclusion : In traditional oriental medicine, hiccup is considered to be caused by uprising stomach gi, whereas, in Sa-am acupuncture, it is considered to be caused by weakness and impurity of lung gi. In Sa-am acupuncture, hiccup is divided into five classes; reverse hiccup (treated with large intestine tonification), wind hiccup (treated with liver tonification), fire hiccup (treated with heart tonification), damp hiccup (treated with spleen tonification) and cold hiccup (treated with kidney tonification). In traditional oriental medicine, hiccup is treated by way of lowering the uprising stomach gi, while, in Sa-am acupuncture, hiccup is treated by way of removing whichever of the original cause of hiccups (impurity of large intestine, damage to liver, dry heat of heart, impairment of spleen, exhaustion of kidney) caused the weakness and the impurity of lung gi. In Sa-am acupuncture, the therapeutic mode for all the five causes of hiccups is tonification mode.