• Korean Journal of Food Science and Technology
• /
• v.47 no.5
• /
• pp.567-573
• /
• 2015

In this study, volatile compounds in nine commercial Japanese distilled liquors (Shochu) were isolated by headspace solid-phase microexrraction (SPME) and analyzed by gas chromatography (GC) and GC-mass spectrometry (MS). A total of 76 volatile components, including 48 esters, 13 alcohols, and 15 miscellaneous components, were identified. Esters and alcohols constituted the largest groups of quantified volatiles. Differences in volatile components among the distilled liquors and possible sample grouping were examined by applying principal component analyses to the GC-MS data sets. The first and second principal components explained 77.92% of the total variation across the samples. The samples using barley koji showed higher overall concentrations of total volatile components. Additionally, the principal component analysis did not reveal any sample grouping based on the raw material used.

• Journal of Broadcast Engineering
• /
• v.26 no.5
• /
• pp.608-620
• /
• 2021

Recent deep learning-based face super-resolution (FSR) works showed significant performances by utilizing facial prior knowledge such as facial landmark and dictionary that reflects structural or semantic characteristics of the human face. However, most of these methods require additional processing time and memory. To solve this issue, this paper propose an efficient FSR models using knowledge distillation techniques. The intermediate features of teacher network which contains dictionary information based on major face regions are transferred to the student through adversarial multi-scale features distillation. Experimental results show that the proposed model is superior to other SR methods, and its effectiveness compare to teacher model.

• 농업기술회보
• /
• v.43 no.5 s.506
• /
• pp.47-53
• /
• 2006

• Journal of Life Science
• /
• v.28 no.12
• /
• pp.1461-1468
• /
• 2018

The Alaska Pollock (Gadus chalcogrammus) is distributed in an arc across the North Pacific Ocean. Distilled water extracts (DWE) and ethanol extracts (ETE) of 1.0 mg/ml concentrations of raw Alaska Pollock roe, premium Gochujang Pollock roe, and premium fermented Pollock roe seasoning were evaluated for estimated 2,2'-azino-bis (3-ethylbenzthiazoline-6-sulphonic acid) (ABTS), ferric reducing antioxidant potential (FRAP), and total phenolic content. The ABTS scavenging activity of the raw Alaska roe DWE and ETE were evaluated at 50.1% and 53.1%, respectively. The ABTS scavenging activity of the Gochujang roe DWE was 68.7% and of the ETE was 70.4%; for the fermented seasoning it was 71.3% and 71.6% for the DWE and ETE, respectively. The ABTS $EC_{50}$ values of the raw roe DWE and ETE were 12.49 ug/ml and 12.21 ug/ml, respectively. The FRAP $EC_{50}$ values of the Gochujang roe DWE and ETE were 10.67 ug/ml and 10.56 ug/ml, respectively, and the $EC_{50}$ values for total phenolic content for the fermented seasoning DWE and ETE were 10.45 ug/ml and 10.31 ug/ml, respectively. When Gallic was acid used as a control, the relative total phenolic content scavenging activity in each ETE was 52.0% (raw Alaska roe), 61.1% (Gochujang roe), and 63.6% (fermented seasoning). In the present study, higher ABTS, FRAP, and total phenolic content were observed in the Gochujang Pollock roe and fermented roe seasoning than in the Alaska Pollock roe.

• Food Science and Preservation
• /
• v.13 no.5
• /
• pp.623-628
• /
• 2006

This study was performed to develop extract materials from Elsholtzia splendens by analyzing the functionality and aroma profile by the simultaneous steam distillation extraction. The qualities of extracts such as total yield, total phenolic compound and electron donation ability were affected by extraction temperature than time. The main flavor compounds were analyzed as estragole, thymol and beta-caryophyllene. The response variables had significant with extraction temperature than tim. The established model was suitable to predict calculated value in experimental ranges. The optimum extinction conditions, which were limited of maximum value for dependent variables were $108^{\circ}C$ and 2.1 hr in the simultaneous steam distillation extraction method.

• Korean Journal of Food Science and Technology
• /
• v.49 no.5
• /
• pp.486-493
• /
• 2017

The purpose of this study was to compare the differences of seven distilled soju samples that were presented at the main stage of the 2015 Korean Liquor Contest. The characteristics of the awarded products and other products were compared. The general components of prize-winning soju showed that the total acidity (0.22%) and alcohol content (38.44%) were higher than other soju (0.10, 22.81%), and that these properties influenced various sensory properties. For the sensory characteristics and consumer preference factors, the descriptive analysis and consumer preference testing were conducted. The descriptive analysis revealed that the (aftertaste) bitterness, corn-silk odor, sake odor/flavor, and overall taste intensity were strongly expressed in the prize-winning distilled soju. For the preference analysis, various preferences, such as "flowery odor/flavor" and "rich odor/flavor" appear to have influenced the positive preferences and "too strong an odor/flavor", "aftertaste bitterness", and "sharp tasting" were found to have a negative effect on preference.

• Smart Media Journal
• /
• v.12 no.9
• /
• pp.89-94
• /
• 2023

Recommender systems provide users with the most favorable items by analyzing explicit or implicit feedback of users on items. Recently, as the size of deep-learning-based models employed in recommender systems has increased, many studies have focused on reducing inference time while maintaining high recommendation accuracy. As one of them, a study on recommender systems with a knowledge distillation (KD) technique is actively conducted. By KD, a small-sized model (i.e., student) is trained through knowledge extracted from a large-sized model (i.e., teacher), and then the trained student is used as a recommendation model. Existing studies on KD for recommender systems have been mainly performed only for implicit feedback settings. Thus, in this paper, we try to investigate the performance and accuracy when applied to explicit feedback settings. To this end, we leveraged a total of five state-of-the-art KD methods and three real-world datasets for recommender systems.

• Plant Journal
• /
• v.13 no.3
• /
• pp.26-29
• /
• 2017

In this research, the composition of the feed stream is optimized in pressure swing distillation for separating of methanol-acetone. It is well known that the composition of feed stream in pressure swing distillation system has a great influence on the feasibility to separate mixture. The workscope of this study is to show better separation efficiency at specified pressure by controlling the composition of feed stream with recycle of two products. Based on the base case without recycle flow, two processes are designed that methanol and acetone are recycled to feed stream, respectively. Each processes are compared with total annual cost and as a result, the base case without recycle flow are most favorable.

• Korean Journal of Food Science and Technology
• /
• v.26 no.5
• /
• pp.649-654
• /
• 1994

In order to provide a quality index of Korean folk sojues, physicochemical properties of Korean folk sojues (Andong soju, Moonbaesul, Leekangju, Jindo Hongju, Chebiwon soju, Yethyang (rice) and Yethyang (barley)), Paekrosul, Chinese kaoliangchiews (Moutaichiew, Ergoutoutiu, Chuyehchingchiew, Zhikukaoliangchiew and Paigal), Japanese Senbatanuki shochu and two whiskies were determined. The pH of sojues ($3.43{\sim}5.85$) were mainly influenced by total acidity which was described as acetic acid. The conductivities of Paekrosul and Leekangju showed relatively high value of $246\;and\;122.7{\mu}S/cm$, respectively. Korean folk sojues and Paekrosul showed maximum absorption at 274 or $278{\sim}280nm$ in the spectrometric study. We performed principal components analysis (PCA) of physicochemical properties and spectrometric data to extract representative characteristics and to compare the similarity or the dissimilarity. The PCA plot showed the distinguished cluster of Korean folk sojues comparing with Chinese kaoliangchiew or Paekrosul etc.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.35 no.10
• /
• pp.1412-1419
• /
• 2006

Concentrations of formaldehyde and acetaldehy de were respectively analysed in forty-five alcoholic beverages obtained from the market. After derivatization with PFBHA, GC-ECD and GC-MSD were employed for analysis. The peak area of aldehyde oximes (derivatives with PFBHA) increased with the increasing ethanol content (5%, 10%, 15%, 20% and 40%). When three-point calibration corves for the selected ethanol concentration (5, 13, 21 and 40%, v/v) were studied, suitable linearity against ethanol concentration was observed only under 5, 13, and 21% (ethanol, v/v). After analysis, maximum content of formaldehyde (average of 0.272 ppm) and acetaldehyde (average of 15.262 ppm) among the observed 45 alcoholic beverages was found from whisk (2 species) while minimum content of formaldehyde (average of 0.009 ppm) and acetaldehyde (average of 0.805 ppm) was found from diluted soju (4 species).