• Progress in Medical Physics
• /
• v.26 no.3
• /
• pp.160-167
• /
• 2015

Radiation exposure from medical diagnostic imaging procedures to patients is one of the most significant interests in diagnostic x-ray system. A miniature x-ray intraoral tube was developed for the first time in the world which can be inserted into the mouth for imaging. Dose evaluation should be carried out in order to utilize such an imaging device for clinical use. In this study, dose evaluation of the new x-ray unit was performed by 1) using a custom made in vivo Pig phantom, 2) determining exposure condition for the clinical use, and 3) measuring patient dose of the new system. On the basis of DRLs (Diagnostic Reference Level) recommended by KDFA (Korea Food & Drug Administration), the ESD (Entrance Skin Dose) and DAP (Dose Area Product) measurements for the new x-ray imaging device were designed and measured. The maximum voltage and current of the x-ray tubes used in this study were 55 kVp, and 300 mA. The active area of the detector was $72{\times}72mm$ with pixel size of $48{\mu}m$. To obtain the operating condition of the new system, pig jaw phantom images showing major tooth-associated tissues, such as clown, pulp cavity were acquired at 1 frame/sec. Changing the beam currents 20 to $80{\mu}A$, x-ray images of 50 frames were obtained for one beam current with optimum x-ray exposure setting. Pig jaw phantom images were acquired from two commercial x-ray imaging units and compared to the new x-ray device: CS 2100, Carestream Dental LLC and EXARO, HIOSSEN, Inc. Their exposure conditions were 60 kV, 7 mA, and 60 kV, 2 mA, respectively. Comparing the new x-ray device and conventional x-ray imaging units, images of the new x-ray device around teeth and their neighboring tissues turn out to be better in spite of its small x-ray field size. ESD of the new x-ray device was measured 1.369 mGy on the beam condition for the best image quality, 0.051 mAs, which is much less than DRLs recommended by IAEA (International Atomic Energy Agency) and KDFA, both. Its dose distribution in the x-ray field size was observed to be uniform with standard deviation of 5~10 %. DAP of the new x-ray device was $82.4mGy*cm^2$ less than DRL established by KDFA even though its x-ray field size was small. This study shows that the new x-ray imaging device offers better in image quality and lower radiation dose compared to the conventional intraoral units. In additions, methods and know-how for studies in x-ray features could be accumulated from this work.

• Journal of Intelligence and Information Systems
• /
• v.23 no.2
• /
• pp.1-17
• /
• 2017

The deep learning framework is software designed to help develop deep learning models. Some of its important functions include "automatic differentiation" and "utilization of GPU". The list of popular deep learning framework includes Caffe (BVLC) and Theano (University of Montreal). And recently, Microsoft's deep learning framework, Microsoft Cognitive Toolkit, was released as open-source license, following Google's Tensorflow a year earlier. The early deep learning frameworks have been developed mainly for research at universities. Beginning with the inception of Tensorflow, however, it seems that companies such as Microsoft and Facebook have started to join the competition of framework development. Given the trend, Google and other companies are expected to continue investing in the deep learning framework to bring forward the initiative in the artificial intelligence business. From this point of view, we think it is a good time to compare some of deep learning frameworks. So we compare three deep learning frameworks which can be used as a Python library. Those are Google's Tensorflow, Microsoft's CNTK, and Theano which is sort of a predecessor of the preceding two. The most common and important function of deep learning frameworks is the ability to perform automatic differentiation. Basically all the mathematical expressions of deep learning models can be represented as computational graphs, which consist of nodes and edges. Partial derivatives on each edge of a computational graph can then be obtained. With the partial derivatives, we can let software compute differentiation of any node with respect to any variable by utilizing chain rule of Calculus. First of all, the convenience of coding is in the order of CNTK, Tensorflow, and Theano. The criterion is simply based on the lengths of the codes and the learning curve and the ease of coding are not the main concern. According to the criteria, Theano was the most difficult to implement with, and CNTK and Tensorflow were somewhat easier. With Tensorflow, we need to define weight variables and biases explicitly. The reason that CNTK and Tensorflow are easier to implement with is that those frameworks provide us with more abstraction than Theano. We, however, need to mention that low-level coding is not always bad. It gives us flexibility of coding. With the low-level coding such as in Theano, we can implement and test any new deep learning models or any new search methods that we can think of. The assessment of the execution speed of each framework is that there is not meaningful difference. According to the experiment, execution speeds of Theano and Tensorflow are very similar, although the experiment was limited to a CNN model. In the case of CNTK, the experimental environment was not maintained as the same. The code written in CNTK has to be run in PC environment without GPU where codes execute as much as 50 times slower than with GPU. But we concluded that the difference of execution speed was within the range of variation caused by the different hardware setup. In this study, we compared three types of deep learning framework: Theano, Tensorflow, and CNTK. According to Wikipedia, there are 12 available deep learning frameworks. And 15 different attributes differentiate each framework. Some of the important attributes would include interface language (Python, C ++, Java, etc.) and the availability of libraries on various deep learning models such as CNN, RNN, DBN, and etc. And if a user implements a large scale deep learning model, it will also be important to support multiple GPU or multiple servers. Also, if you are learning the deep learning model, it would also be important if there are enough examples and references.

• Journal of Trauma and Injury
• /
• v.23 no.2
• /
• pp.102-106
• /
• 2010

Purpose: With the increasing numbers of foreign residents in Korea, the need for an emergency medical care system for foreign patients seems to be growing. Sometimes, a foreigner admitted to an emergency room is not treated sufficiently due to the absence of insurance, facility in the Korean language, and a guardian. The management of a foreigner with trauma in the ER is difficult due to various problems such as social and economic status. The purpose of this study was to investigate the current management status of foreigners with penetrating wounds in the emergency room. Methods: This study is an analysis of 580 patients that were diagnosed with penetrating wounds in one teaching Hospital from Jan. 1, 2008 to Dec. 31, 2008. We analyzed results according to nationality, alcohol ingestion, intentional or accidental trauma, trauma mechanism, injury severity, management time in the ER, and outcome in the ER. Results: Of the total 580 patients, 486 patients (83.8%) were native Koreans and 94 patients (16.2%) were foreigners. According to the Revised Trauma Score, the average score of native Korean patients was 7.808, and the average score of foreign patients was 7.638. Of native Korean patients, 22.6% had knife wounds while 38.3% of foreign patients did. Of native Korean patients, 17.3% experienced intentional trauma while 33.0% of the foreign patients did. Of native Korean patients, 22.5% had ingested alcohol while 49.4% of the foreigners had. Of native Korean patients, 10.5% were admitted while 7.6% of the foreign patients were. Of native Korean patients, 14.2% were discharged against medical advice (DAMA), while 18.5% of foreign patients were. Of native Korean patients, 1.2% ran away while 8.7% of the foreign patients did. Conclusion: Stabbing was the most common cause of penetrating wounds in foreigner patients in this study. Intentional trauma was more common in foreigners with penetrating wounds than in native Koreans. The severity was higher in foreigners with penetrating wounds than it was in native Koreans, and patients who ran away or were discharged against medical advice were more commonly foreigners with penetrating wounds. Social insurance or policy is needed for the management of foreigners with penetrating wounds.

• Journal of KIISE:Computing Practices and Letters
• /
• v.5 no.4
• /
• pp.416-425
• /
• 1999

복잡한 무결성 제약 조건을 효율적으로 확인하기 위해 제약 조건들을 룰 베이스(rule base)에 저장하고 별도의 룰 관리 시스템과 제약 조건 관리 시스템을 통해 제약 조건을 확인하는 기법이 많은 연구자들에 의해 연구되고 발표되었다. 그러나 제약 조건 관리 시스템이 실행시간에 응용 프로그램을 항상 모니터링하고 있다가 데이타의 수정이 요청될 때마다 개입하여 프로세스를 중단시키고 관련 제약 조건을 확인하는 기존의 방법들은 처리 시간의 지연을 피할 수 없다. 본 논문은 컴파일 시간에 제약 조건 확인 코드를 응용 프로그램에 미리 삽입할 것을 제안한다. 응용 프로그램 자체 내에 제약 조건 확인 코드가 삽입되기 때문에 실행 시간에 다른 시스템의 제어를 받지 않고 직접 제약 조건의 확인 및 데이타베이스의 접근이 가능해져서 처리 시간의 지연을 피할 수 있을 것이다. 이를 위해 어떤 구문이 제약 조건의 확인을 유발하는 지를 추적하였고, 컴파일러가 그러한 구문을 어떻게 전처리 과정에서 검색하는지 그리고 그러한 구문마다 어떻게 해당 제약 조건 확인 코드를 삽입할 수 있는 지를 객체지향1) 데이타베이스 언어인 Objectivity/C++에 대해 gcc의 YACC 코드를 변경함으로써 보여 주었다.Abstract To cope with the complexity of handling integrity constraints, numerous researchers have suggested to use a rule-based system, where integrity constraints are expressed as rules and stored in a rule base. A rule manager and an integrity constraint manager cooperate to check the integrity constraints efficiently. In this approach, however, the integrity constraint manager has to monitor the activity of an application program constantly to catch any database operation. For each database operation, it has to check relevant rules with the help of the rule manager, resulting in considerable delays in database access. We propose to insert the constraints checking code in the application program directly at compile time. With constraints checking code inserted, the application program can check integrity constraints by itself without the intervention of the integrity constraint manager. We investigate what kind of statements require the checking of constraints, show how the compiler can detect those statements, and show how constraints checking code can be inserted into the program, by modifying the GCC YACC file for Objectivity/C++, an object-oriented database programming language.

• The Korean Journal of Archival Studies
• /
• no.64
• /
• pp.27-62
• /
• 2020

The article has the characteristics of a preliminary writing about how to look at the trend of new archives 'fever' and 'impulsion' emerging around the domestic and foreign art world, which have not been paid much attention yet in the 'mainstream' archive research, and how to accept it independently. Specifically, this study aims to examine how archival art is involved in history and memory with aesthetic attitudes and methods through observation of recent tendency of domestic archive art, and what implications or influence the 'archival impulse' phenomenon in the art world can have on the research trend of 'archival studies.' First, I would like to look at the meaningful movement to reinterpret and actively accept archival impulses in concrete overseas cases, that is, the archive system of a public archive in the United States. This is followed by an attempt to explore the characteristics and characteristics of creative works that are carried out through the medium of archives, that has not yet reached the level of organization of specific archive methods but are sporadically attempted in the domestic art world. It examines how so-called 'archive artists' record unrecorded in a way that is not observed in the existing archival world, and how they summon and include excluded history in aesthetic language. In conclusion, this study explores the possibility of pulling the historical records of tradition out from archival boxes and reinterpreting them as living archives within the contemporary emotional structure from this new artistic trend called 'archival art'.

• Journal of the Korea Society for Simulation
• /
• v.29 no.4
• /
• pp.33-46
• /
• 2020

The purpose of Electronic warfare is to disturbe, neutralize, attack, and destroy the opponent's electronic warfare weapon system or equipment. Suppression of Enemy Air Defense (SEAD) mission is aimed at incapacitating, destroying, or temporarily deteriorating air defense networks such as enemy surface-to-air missiles (SAMs), which is a representative mission supported by electronic warfare. This paper develops a simulator for analyzing the effectiveness of SEAD missions under electronic warfare support using C++ language based on the DEVS (Discrete Event Systems Specification) model, the usefulness of which has been proved through case analysis with examples. The SEAD mission of the friendly forces is carried out in parallel with SSJ (Self Screening Jamming) electronic warfare under the support of SOJ (Stand Off Jamming) electronic warfare. The mission is assumed to be done after penetrating into the enemy area and firing HARM (High Speed Anti Radiation Missile). SAM response is assumed to comply mission under the degraded performance due to the electronic interference of the friendly SSJ and SOJ. The developed simulator allows various combinations of electronic warfare equipment specifications (parameters) and operational tactics (parameters or algorithms) to be input for the purpose of analysis of the effect of these combinations on the mission effectiveness.

• Journal of Korea Entertainment Industry Association
• /
• v.13 no.8
• /
• pp.151-156
• /
• 2019

In recent years, the Korean realistic theme of the film momentum gradually rising. Realistic films do not stick to the business and market, and do not simply cater to the audience's needs for watching movies. They reflect social violence and cruel reality, allowing the audience to observe the structural contradictions in reality and think about the direction when watching movies. At the recent cannes film festival, "parasite" won the top prize palm in cannes by an overwhelming margin, with the highest score of 3.3 issues. Although this film is positioned as a thriller with comedy elements, it presents the opposite life images of Korean classes to the audience in a parasitic way, which not only expands the possibility and artistry of realistic film aesthetics, but also enhances the appreciation of the film and gives play to its own aesthetic value. Focusing on the technical and literary nature of the film, and having a high degree of attention to real life, it is an excellent work that tells about class opposition and thinking about reality. This paper considers and analyzes the content, form and creation method of parasite, and discusses the continuous exploration and attempt of realistic film to image language under the demand of market and system, evolving into new aesthetic expression.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.360-363
• /
• 2022

According to the 2018 Kaggle ML & DS Survey, among the proportions of frameworks for machine learning and data science, TensorFlow and Keras each account for 41.82%. It was found to be 34.09%, and in the case of development programming, it is confirmed that about 82% use Python. A significant number of machine learning and deep learning structures utilize the Keras framework and Python, but in the case of Python, distribution and execution are limited to the Python script environment due to the script language, so it is judged that it is difficult to operate in various environments. This paper implemented a machine learning and deep learning system using C# and Keras running in Visual Studio 2019. Using the Mnist dataset, 100 tests were performed in Python 3.8,2 and C# .NET 5.0 environments, and the minimum time for Python was 1.86 seconds, the maximum time was 2.38 seconds, and the average time was 1.98 seconds. Time 1.78 seconds, maximum time 2.11 seconds, average time 1.85 seconds, total time 37.02 seconds. As a result of the experiment, the performance of C# improved by about 6% compared to Python, and it is expected that the utilization will be high because executable files can be extracted.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• The Critical Review of Religion and Culture
• /
• no.31
• /
• pp.255-287
• /
• 2017

This paper analyzes the religious circuit of suicidal concept based on verbal expression and ritual acts, which are found in the suicide discourse of Korean Catholic Church and Protestant Church. In the relationship of suicide and religion, it is easily overlooked the religious circuit and its construction that forms the concept of suicide among the religious laymen. It is assumed that the belief system of traditional religions prohibits suicide and the laymen accordingly construct a perception or concept of suicide along with this belief system. Various studies on this subject have proved it. However, in order to understand the religious way of constructing the concept of suicide on a personal level, it is necessary to pay attention to the religious environment in which the concepts and emotions of suicide circulate. The laymen do not passively and perfectly accept the finely established suicide concept provided by the doctrine or the theology. Rather, the laymen tend to collect the pieces of concept over the suicide that are drifting in the religious environment of his/her daily routine life and to make an concept of suicide in an incomplete form. We can find the unstable and imperfect traits of such a suicide concept through the experience of suicide survivors who have a religious background. For the suicide survivors with religious beliefs, they resist the formal doctrinal and theological provisions to suicide, or try to understand the notion of suicide in their own contexts. In terms of linguistic expressions and ritual acts relating to suicide, the attentions are differently directed in the public and the private domain among the religious groups. Considering on the high rates of suicide in Korean society, the Korean Catholic Churches are increasingly tolerant over the suicide and accept it in the public sphere. It is unlikely when comparing to the negative attitudes of the suicide in the past. However, such tolerance does not go beyond the doctrinal and ethical judgment that defines suicide as a serious sin. The once-committed lay believer's speech and gestures usually contain the various emotions, such as sadness, grief, anxiety, regretfulness, eagerness, and pain in the private spheres. The language and gestures with these emotions have been activated in the religious circuits of suicide, being extended to the religious apparatus for the person who died of suicide. In case of Protestantism, the institutional organizations, such as the particular denominations and the individual-churchism of the Korean Protestant Churches, and their own interpretations of the Bible have in the private sphere strongly effected on the linguistic expressions and the rituals related to the suicide. The religious-ethical judgment of the suicide is varied how the suicide is interpreted by the theologians and the pastors. And the ritual acts for healing the complex feelings and the psychological wounds of the suicide survivors are not actively explored and adopted yet. It makes harder to approach and heal the protestant followers since they emphasize the innermost belief and the salvation assurance faith.