• Journal of the Korean Society of Clothing and Textiles
• /
• v.46 no.1
• /
• pp.165-181
• /
• 2022

The fashion industry is creating innovative business models using artificial intelligence. To efficiently utilize artificial intelligence (AI), fashion data must be classified. Until now, such data have been classified focusing only on the objective properties of fashion products. Their subjective attributes, such as fashion brand sensibilities, are holistic and heuristic intuitions created by a combination of design elements. This study aims to improve the performance of collaborative filtering in the fashion industry by extracting fashion brand sensibility using computer vision technology. The image data set of fashion brand sensibility consists of high-end fashion brand photos that share sensibilities and communicate well in fashion. About 26,000 fashion photos of 11 high-end fashion brand sensibility labels have been collected from the 16FW to 21SS runway and 50 years of US Vogue magazines beginning from 1971. We use EfficientNet-B1 to establish the main architecture and fine-tune the network with ImageNet-ILSVRC. After training fashion brand sensibilities through deep learning, the proposed model achieved an F-1 score of 74% on accuracy tests. Furthermore, as a result of comparing AI machine and human experts, the proposed model is expected to be expanded to mass fashion brands.

• International Journal of Computer Science & Network Security
• /
• v.22 no.5
• /
• pp.175-181
• /
• 2022

COVID-19 is an acute respiratory syndrome that affects the host's breathing and respiratory system. The novel disease's first case was reported in 2019 and has created a state of emergency in the whole world and declared a global pandemic within months after the first case. The disease created elements of socioeconomic crisis globally. The emergency has made it imperative for professionals to take the necessary measures to make early diagnoses of the disease. The conventional diagnosis for COVID-19 is through Polymerase Chain Reaction (PCR) testing. However, in a lot of rural societies, these tests are not available or take a lot of time to provide results. Hence, we propose a COVID-19 classification system by means of machine learning and transfer learning models. The proposed approach identifies individuals with COVID-19 and distinguishes them from those who are healthy with the help of Deep Visual Embeddings (DVE). Five state-of-the-art models: VGG-19, ResNet50, Inceptionv3, MobileNetv3, and EfficientNetB7, were used in this study along with five different pooling schemes to perform deep feature extraction. In addition, the features are normalized using standard scaling, and 4-fold cross-validation is used to validate the performance over multiple versions of the validation data. The best results of 88.86% UAR, 88.27% Specificity, 89.44% Sensitivity, 88.62% Accuracy, 89.06% Precision, and 87.52% F1-score were obtained using ResNet-50 with Average Pooling and Logistic regression with class weight as the classifier.

• Korea Science and Art Forum
• /
• v.34
• /
• pp.275-290
• /
• 2018

The study began with the recognition that the animations movie need objective and reasonable methods to classify conflicts in visual to analyze conflicts centering on narratives. Study the emotions of the hero in conflict. The purpose of the study is to analyze conflict intensity and emotion. The results and contents of the study are as follows. First, we found a Turning Point and suggested a conflict classification model (Conflict 6B Model). Second, Based on the conflict classification model, the conflict based shot DB was extracted. Third, I found strength and emotion in inner and super personal conflicts. Fourth, Experiments and tests of strength and emotion were conducted in internal and super personal conflicts. The results of this study are metadata extracted from the emotional research on conflict. It is expected to be applied to video indexing of conflicts.

• The Journal of the Acoustical Society of Korea
• /
• v.26 no.2
• /
• pp.87-94
• /
• 2007

In active sonar system, aspect angle and length of a target can be estimated by calculating the cross-correlation between left and right split-beams of a LFM(Linear Frequency Modulated) signal. However, high-resolution performances in bearing and range are required to estimate the information of a remote target. Because a certain higher sampling frequency than the Nyquist sampling frequency is required in this performance, an over-sampling process through interpolation method should be required. However, real-time implementation of split-beam processing with over-sampled split-beam outputs on a COTS(commercial off-the-shelf) DSP platform limits its performance because of given throughput and memory capacity. This paper proposes a cumulative processing algorithm for split-beam processing to solve the problems. The performance of the proposed method was verified through some simulation tests. Also, the proposed method was implemented as a real-time system using an ADSP-TS101.

• Journal of the Korea Concrete Institute
• /
• v.23 no.3
• /
• pp.311-320
• /
• 2011

Fly ash (FA), byproduct from power plant has been actively used as mineral admixture for concrete. However, since bottom ash (BA) is usually used for land reclaim or subbase material, more active reuse plan is needed. Pond ash (PA) obtained from reclaimed land is mixed with both FA and BA. In this study, 6 PA from different domestic power plant are prepared and 5 different replacement ratios (10%, 20%, 30%, 50%, and 70%) for fine aggregate substitutes are considered to evaluate engineering properties of PA as fine aggregate and durability performance of PA concrete. Tests for fine aggregate of PA for fineness modulus, density and absorption, soundness, chloride and toxicity content, and alkali aggregate reaction are performed. For PA concrete, durability tests for compressive strength, drying shrinkage, chloride penetration/diffusion, accelerated carbonation, and freezing/thawing are performed. Also, basic tests for fresh concrete like slump and air content are performed. Although PA has lower density and higher absorption, its potential as a replacement material for fine aggregate is promising. PA concrete shows a reasonable durability performance with higher strength with higher replacement ratio. Finally, best PA among 6 samples is selected through quantitative classification, and limitation of PA concrete application is understood based on the test results. Various tests for engineering properties of PA and PA concrete are discussed in this paper to evaluate its application to concrete structure.

• The Korean Journal of Pesticide Science
• /
• v.12 no.2
• /
• pp.141-147
• /
• 2008

Globally Harmonized System (GHS) for classification and labelling will provide an internationally agreed hazard classification system of chemical products and for communication of those hazards. This study aimed for establishment of the oral acute toxic class (ATC) method and application of GHS on pesticides. The ATC method was developed for determining $LD_{50}$ estimates of chemical substances with significantly fewer animals than needed when applying conventional $LD_{50}$ tests. We carried out $LD_{50}$ test and ATC test on 13 pesticides, Although methidathion EC and parathion-ethyl EC showed significantly different in $LD_{50}$ values between $LD_{50}$ test and ATC method, there are no difference in toxicity class by GHS. Both tests on the other pesticides showed almost equal results and toxicity class by GHS. Therefore, this study indicated high possibility of application of ATC method and GHS on pesticides.

• Journal of the Korean Wood Science and Technology
• /
• v.36 no.1
• /
• pp.36-44
• /
• 2008

It is necessary to assess wood quality and strength of the clear specimens from used members in old wooden structures for the reuse of old members on the same structure or a new structure. Wood species classification by microscope observation of each wood member that was used in truss cord and temple, and several physical and strength tests by the specification of present KS standards were conducted to compare with some references. From the comparison of strengths with references, Korean larch gives relatively better wood quality and mechanical properties than other wood species. No significant deterioration of cell wall was found by microscopic observation for the sound wood part that was selected visually. Tensile specimens with 3 mm in thickness on the middle span showed greater strength than 5 mm thick specimens, which explains that dimension of tensile specimen should be examined for evaluating precise tensile strength properties. Other tests, compression, shear, and bending, are adoptable for each strength properties. Test methods for the evaluation of basic strengths and fastener connections for old wood species should be further examined.

• International Journal of Contents
• /
• v.10 no.3
• /
• pp.1-8
• /
• 2014

In this paper, we collect and analyze the status of mobile phone applications (hereafter apps) in the healthcare and fitness category of the Apple iTunes App Store and Google Play Store. We determine the number of apps and analyze statistical aspects such as classifications, age rating, fees, and user evaluation of the popular items. As of September 30, 2013, there were 236 popular apps available from iTunes. Google Play offered 720 apps. We discover that apps for healthcare and fitness are diverse. Apps for physical exercise have the greatest popularity. The proportions of apps that are suitable for all ages among the Google and iTunes popular apps are 55.8% and 89.4%, respectively. The user evaluation of apps in iTunes is relatively less positive. We determine that the proportion of paid apps to free apps in Google is higher than that of the apps in iTunes. We perform hypothesis tests and find statistically significant differences in age rating and perceived satisfaction between the apps of the Apple iTunes App Store and Google Play Store. However, we find no meaningful differences in the classification and price of the apps between the two app stores. We perform hypothesis tests to verify the differences in age rating and perceived satisfaction between the paid and free apps within and across the Google Play Store and iTunes App Store. There are statistically significant differences in the age rating between the paid and free apps in the Google play store, between the Google free and iTunes free apps, between the Google paid and iTunes paid apps, between the Google free and iTunes paid apps, and between the Google paid and iTunes free apps. There are statistically significant differences in the perceived satisfaction between the Google free and iTunes free apps, between the Google paid and iTunes paid apps, between the Google free and iTunes paid apps, and between the Google paid and iTunes free apps.

• Journal of Advanced Navigation Technology
• /
• v.18 no.3
• /
• pp.209-214
• /
• 2014

Current black boxes have a problem of storing unnecessary imagery data recordings without data classification. For this reason, users have to erase videos every time. This method is inadequate for black boxes with limited memory capacity. In this paper, we design and implement a system that recognizes traffic accident situations and saves these recordings by classifying them according to weighted values. The system was made to save video recorded at a 30-sec interval of every event to black box folders by changing names based on weighted value data under the external environment in a 1:10 scale model car. Based on this, when the tests were performed as a major car accident while driving, the videos were created in w2 folder, and when the tests were performed as a minor car accident while stopped, the videos were created in w1 folder.

• Journal of the Korean Society for Railway
• /
• v.11 no.1
• /
• pp.54-60
• /
• 2008

In the trackbed design using elastic multilayer model, the stress-dependent resilient modulus $(E_R)$ is an important input parameter, that is, reflects substructure performance under repeated traffic loading. However, the evaluation method for resilient modulus using repeated loading triaxial test is not fully developed for practical purpose, because of costly equipment and the significantly fluctuated values depending on the testing equipment and laboratory personnel. The this study, the paper will present an indirect method to estimate the resilient modulus using dynamic properties. The resilient modulus of crushed stone, which is the typical material of sub-ballast, was calculated with the measured dynamic properties and the range of stress level of the sub-ballast, and approximated with the power model combined with bulk and deviatoric stresses. The resilient modulus of coarse grained material decreases with increasing deviatoric stress at a confining pressure, and increases with increasing bulk stress. Sandy soil (SM classified from Unified Soil Classification System) of subgrade was also evaluated and best fitted with the power model of deviatoric stress only.