• Computers and Concrete
• /
• v.23 no.6
• /
• pp.421-431
• /
• 2019

The alarming rate of depletion of natural stone based coarse aggregates is a cause of great concern. The coarse aggregates occupy nearly 60-70% by volume of concrete being produced. Research efforts are on to look for alternatives to stone based coarse aggregates from sustainability point of view. Response surface methodology (RSM) is adopted to study and address the effect of ferrochrome slag (FCS) replacement to coarse aggregate replacement in the ordinary Portland cement (OPC) based concretes. RSM involves three different factors (ground granulated blast furnace slag (GGBS) as binder, flyash (FA) as binder, and FCS as coarse aggregate), with three different levels (GGBS (0, 15, and 30%), FA (0, 15, and 30%) and FCS (0, 50, and 100%)). Experiments were carried out to measure the responses like, workability, density, and compressive strength of FCS based concretes. In order to optimize FCS replacement in the OPC based concretes, three different traditional optimization techniques were used (grey relational analysis (GRA), technique for order of preference by similarity (TOPSIS), and desirability function approach (DFA)). Traditional optimization techniques were accompanied with principal component analysis (PCA) to calculate the weightage of responses measured to arrive at the final ranking of replacement levels of GGBS, FA, and FCS in OPC based concretes. Hybrid combination of PCA-TOPSIS technique is found to be significant when compared to other techniques used. 30% GGBS and 50% FCS replacement in OPC based concrete was arrived at, to be optimal.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1597-1610
• /
• 2022

At present, the main method of high-speed train chassis detection is using computer vision technology to extract keypoints from two related chassis images firstly, then matching these keypoints to find the pixel-level correspondence between these two images, finally, detection and other steps are performed. The quality and accuracy of image matching are very important for subsequent defect detection. Current traditional matching methods are difficult to meet the actual requirements for the generalization of complex scenes such as weather, illumination, and seasonal changes. Therefore, it is of great significance to study the high-speed train image matching method based on deep learning. This paper establishes a high-speed train chassis image matching dataset, including random perspective changes and optical distortion, to simulate the changes in the actual working environment of the high-speed rail system as much as possible. This work designs a convolutional neural network to intensively extract keypoints, so as to alleviate the problems of current methods. With multi-level features, on the one hand, the network restores low-level details, thereby improving the localization accuracy of keypoints, on the other hand, the network can generate robust keypoint descriptors. Detailed experiments show the huge improvement of the proposed network over traditional methods.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.313-315
• /
• 2021

Timely and accurate diagnosis of lung diseases using Chest X-ray images has been gained much attention from the computer vision and medical imaging communities. Although previous studies have presented the capability of deep convolutional neural networks by achieving competitive binary classification results, their models were seemingly unreliable to effectively distinguish multiple disease groups using a large number of x-ray images. In this paper, we aim to build an advanced approach, so-called Ensemble Knowledge Distillation (EKD), to significantly boost the classification accuracies, compared to traditional KD methods by distilling knowledge from a cumbersome teacher model into an ensemble of lightweight student models with parallel branches trained with ground truth labels. Therefore, learning features at different branches of the student models could enable the network to learn diverse patterns and improve the qualify of final predictions through an ensemble learning solution. Although we observed that experiments on the well-established ChestX-ray14 dataset showed the classification improvements of traditional KD compared to the base transfer learning approach, the EKD performance would be expected to potentially enhance classification accuracy and model generalization, especially in situations of the imbalanced dataset and the interdependency of 14 weakly annotated thorax diseases.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.1
• /
• pp.1-14
• /
• 2024

Time-series forecasting is extensively used in the actual world. Recent research has shown that Transformers with a self-attention mechanism at their core exhibit better performance when dealing with such problems. However, most of the existing Transformer models used for time series prediction use the traditional encoder-decoder architecture, which is complex and leads to low model processing efficiency, thus limiting the ability to mine deep time dependencies by increasing model depth. Secondly, the secondary computational complexity of the self-attention mechanism also increases computational overhead and reduces processing efficiency. To address these issues, the paper designs an efficient multi-layer attention-based time-series forecasting model. This model has the following characteristics: (i) It abandons the traditional encoder-decoder based Transformer architecture and constructs a time series prediction model based on multi-layer attention mechanism, improving the model's ability to mine deep time dependencies. (ii) A cross attention module based on cross attention mechanism was designed to enhance information exchange between historical and predictive sequences. (iii) Applying a recently proposed sparse attention mechanism to our model reduces computational overhead and improves processing efficiency. Experiments on multiple datasets have shown that our model can significantly increase the performance of current advanced Transformer methods in time series forecasting, including LogTrans, Reformer, and Informer.

• Journal of Conservation Science
• /
• v.27 no.4
• /
• pp.421-430
• /
• 2011

Thermal conditions in practice at the traditional bronze workshop of the Korean Folk Village in Yongin were examined along with the microstructures of some high tin bronze objects made there. Laboratory experiments approximating the conditions of the workshop were also carried out and the results were compared. The operating temperature of the workshop furnace was measured to range from $750^{\circ}C$ to $850^{\circ}C$ while the surface temperature of an object, upon its removal from the furnace for additional thermo-mechanical treatments, was generally in the range of $600^{\circ}C$ to $685^{\circ}C$. This variation in working temperatures was reflected in varying microstructures developed upon quenching. The products of the Folk Village were found to consist of microstructures where the ${\alpha}$ grains of the Cu-Sn system were distributed in the background of different phases including the ${\beta}$-martensite phase, retained ${\gamma}$ phase, ${\alpha}+{\delta}$ eutectoid or their mixtures. This variability, which is also identified in objects made in ancient times as well as in our laboratory experiments, suggests that the actual thermal conditions given during the quenching treatments are much more complicated than is inferred from the temperature measurements. This paper will present detailed accounts of the thermo-mechanical treatments as observed in the high tin bronze workshop of the Korean Folk Village and discuss the evolution of varying microstructures in terms of the substantial variability involved in the implementation of the traditional forged high tin bronze technology of Korea.

• CRM연구
• /
• v.4 no.2
• /
• pp.19-29
• /
• 2011

There is a large difference between purchasing patterns in an online shopping mall and in an offline market. This difference may be caused mainly by the difference in accessibility of online and offline markets. It means that an interval between the initial purchasing decision and its realization appears to be relatively short in an online shopping mall, because a customer can make an order immediately. Because of the short interval between a purchasing decision and its realization, an online shopping mall transaction usually contains fewer items than that of an offline market. In an offline market, customers usually keep some items in mind and buy them all at once a few days after deciding to buy them, instead of buying each item individually and immediately. On the contrary, more than 70% of online shopping mall transactions contain only one item. This statistic implies that traditional data mining techniques cannot be directly applied to online market analysis, because hardly any association rules can survive with an acceptable level of Support because of too many Null Transactions. Most market basket analyses on online shopping mall transactions, therefore, have been performed by expanding the co-occurrence criteria of traditional association rule mining. While the traditional co-occurrence criteria defines items purchased in one transaction as concurrently purchased items, the expanded co-occurrence criteria regards items purchased by a customer during some predefined period (e.g., a day) as concurrently purchased items. In studies using expanded co-occurrence criteria, however, the criteria has been defined arbitrarily by researchers without any theoretical grounds or agreement. The lack of clear grounds of adopting a certain co-occurrence criteria degrades the reliability of the analytical results. Moreover, it is hard to derive new meaningful findings by combining the outcomes of previous individual studies. In this paper, we attempt to compare expanded co-occurrence criteria and propose a guideline for selecting an appropriate one. First of all, we compare the accuracy of association rules discovered according to various co-occurrence criteria. By doing this experiment we expect that we can provide a guideline for selecting appropriate co-occurrence criteria that corresponds to the purpose of the analysis. Additionally, we will perform similar experiments with several groups of customers that are segmented by each customer's average duration between orders. By this experiment, we attempt to discover the relationship between the optimal co-occurrence criteria and the customer's average duration between orders. Finally, by a series of experiments, we expect that we can provide basic guidelines for developing customized recommendation systems.

• Journal of Information Processing Systems
• /
• v.14 no.5
• /
• pp.1102-1113
• /
• 2018

Vibrotactile feedback is widely used in designing non-visual interactions on mobile phones, such as message notification, non-visual reading, and blind use. In this work, novel vibrotactile codes are presented to implement a non-visual text reading system for mobile phones. The 26 letters of the English alphabet are formed in an index table with four rows and seven columns, and each letter is mapped using the codes of vibrations. Two kinds of vibrotactile codes are designed with the actuator's on and off states and with specific lengths (short and long) assigned to each state. To improve the efficiency of tactile perception and user satisfaction, three user experiments are conducted. The first experiment explores the maximum number of continuous vibrations and minimum vibration time of the actuator's on and off states that the human can perceive. The second experiment determines the minimum interval between continuous vibrations. The vibrotactile reading system is designed and evaluated in the third experiment according to the results of the two preceding experiments. Results show that the character reading accuracy reaches 91.7% and the character reading speed is approximately 617.8 ms. Our method has better reading efficiency and is easier to learn than the traditional Braille coding method.

• Fisheries and Aquatic Sciences
• /
• v.6 no.1
• /
• pp.34-40
• /
• 2003

Three panel trammel nets were made illegal for the brown shrimp (Penaeus japonicus) or fishes by Korean fisheries law while trammel nets for fleshy prawn in the West sea remained legal. In this study a single panel tangle net with vertical loop lines rigged between the float line and sinker line was specially designed to catch brown shrimp. This net was developed for the first time after observation of the brown shrimp behaviour when reacting to a net in an observation tank. In field experiments these single tangle nets were compared with the traditional trammel nets in the coastal waters of the Keoje area. The mean number of the brown shrimp for 53 fishing operations was 1.13 per unit panel of the single tangle nets when fitted with the vertical loop lines. This was $84\%$ of the mean catch of 1.36 achieved with the trammel nets. These results of fishing experiments using single tangle nets in the field revealed a high fishing efficiency for the brown shrimp and showed little difference from trammel nets. The size of the brown shrimp or number of by-catch was not different between single tangle nets and trammel nets.

• Journal of Physiology & Pathology in Korean Medicine
• /
• v.24 no.4
• /
• pp.559-565
• /
• 2010

Cancer is already a well-recognized main cause of mortality and the incidence of cancer is increasing steadily. Because conventional treatment modalities for cancer accompanies severe side effects, traditional medicine has been considered as alternatives to reduce the adverse effects and its use has continued to rise in cancer therapy. This study aims to summarize and make a reference of radioprotective effects of herbs worldwide. In this process, this review surveyed all papers of radioprotective-focused studies using herbal medicine in PubMed database and finally 44 papers were included. The type of materials, formation of experiments, type of herbal medicine, their action and mechanisms, and type of cancer were analyzed. The number of studies on radioprotective effects of herbal medicine has increased since 2000. The main formation of experiments was clinical study and the portion was 45% and the proportion of the research using prescriptions was 51% and the research using herbal products was 25%. Herbs and prescriptions having the effects of tonifying and nourishment were used the most. Most of herbal medicine in this study can enhance immune function, increase anti-oxidant effect, regulate cell cycle and increase sensitivity to radiotherapy. This study will provide the useful information on development of herbal medicine having radioprotective effects.

• Rapid Communication in Photoscience
• /
• v.4 no.1
• /
• pp.1-8
• /
• 2015

For understanding molecular mechanisms of photochemical reactions, in particular reactions of proteins with biological functions, it is important to elucidate both the initial reactions from the photoexcited states and the series of subsequent chemical reactions, e.g., conformation, intermolecular interactions (hydrogen bonding, hydrophobic interactions), and inter-protein interactions (oligomer formation, dissociation reactions). Although time-resolved detection of such dynamics is essential, these dynamics have been very difficult to track by traditional spectroscopic techniques. Here, relatively new approaches for probing the dynamics of protein photochemical reactions using time-resolved transient grating (TG) are reviewed. By using this method, a variety of spectrally silent dynamics can be detected and such data provide a valuable description about the reaction scheme. Herein, a blue light sensor protein TePixD is the exemplar. The initial photochemistry for TePixD occurs around the chromophore and is detected readily by light absorption, but subsequent reactions are spectrally silent. The TG experiments revealed conformational changes and changes in inter-protein interactions, which are essential for TePixD function. The TG experiments also showed the importance of fluctuations of the intermediates as the driving force of the reaction. This technique is complementary to optical absorption detection methods. The TG signal contains a variety of unique information, which is difficult to obtain by other methods. The advantages and methods for signal analyses are described in detail in this review.