확장성을 고려한 UCG 저작도구의 구조 설계

성연식,조경은,엄기현 한국게임학회 2008 한국게임학회 논문지 Vol.8 No.4

UCC(User Created Contents) 사이트가 활성화되면서 제작 가능한 콘텐츠 종류가 다양해졌다. 제작 가능한 콘텐츠 중에서 UCG(User Created Game)는 사용자가 직접 제작하고 공유하는 게임이다. UCG의 제작 방법은 사용자의 숙련도에 따라서 두 가지로 분류해 볼 수 있다. 첫 번째, 저작도구를 사용하는 방법은 게임 제작에 익숙하지 않은 초보 사용자에게 적합하지만 사용방법이 쉬워지면서 기능이 제약적이다. 두 번째, 프로그램 언어를 이용한 개발은 게임 제작에 익숙한 고급 사용자에게 적합하지만 개발기간이 길어진다. 이 논문에서는 스크립트를 사용해서 초보 사용자에게 적합한 저작도구의 기능을 확장하는 방법을 제안한다. 그리고 제안한 저작도구는 프로그램 소스를 생성하기 때문에 고급 사용자는 제공하지 않는 기능만 추가 구현해서 개발 기간을 단축한다. 제안한 방법의 UCG 제작 과정을 검증하기 위해서 게임 제작에 필요한 GUI, 스크립트 생성 기능과 프로그램 소스 생성 기능을 포함하는 프레임워크를 설계한다. 그리고 설계한 게임 저작도구를 구현해서 개발된 게임 제작 사례를 보임으로써 본 연구에서 제안하는 방법을 소개한다. For the activation and wide distribution of User Created Contents (UCC) sites have diversified user-created possibilities. Among these contents, User Created Game (UCG) sites are places where users can create and share their game contents with others. The method of UCG development can be classified into two categories according to the uses' level of professionalism. First, the method of using the authoring tools is suggested for those unfamiliar with the development or creation of contents. Although the authoring tool is easy to use, there are many functional limitations. Second, development using program languages is suggested for trained advanced users but has the limitation of a prolonged development period. This paper proposes a new method, generating the script which will expand the current functional limitations entailed behind the authoring tools used by first time and less trained users. In order to verify the proposed method in a real UCG development environment, a framework encompassing GUI, script generating function and program source generating function were constructed.

뇌파를 BCI 게임 제어에 활용하기 위한 정규화 방법

성연식,조경은,엄기현 한국게임학회 2010 한국게임학회 논문지 Vol.10 No.6

뇌파는 초기에 원숭이가 모터로 팔을 조작하기 위한 방법에 관한 연구로 시작되었다. 최근에는, 측정한 뇌파를 치매 환자의 치매 진행 정도를 늦추거나 집중력결핍 과잉행동장애 아이들의 집중력을 높이기 위한 연구들이 진행되고 있다. 그리고 저가의 뇌파 측정 장치가 출시되면서 게임 인터페이스로도 사용된다. 뇌파로 게임을 제어할 때의 문제점은 사람마다 평균 진폭, 평균 파장 그리고 평균 진동 횟수가 다르다는 것이다. 뇌파 차이는 뇌파로 게임을 제어할 때 형평성 문제를 발생시키기 때문에 뇌파를 정규화해서 사용하는 방법이 필요하다. 이 논문에서는 정규분포를 사용해서 측정한 뇌파를 정규화하고 제어로 사용할 파형을 계산하는 방법을 제안한다. 이를 위해서 뇌파 변환 과정을 7단계로 나누어 처리하는 프레임워크를 제안하고 각 단계에 필요한 계산과정을 기술한다. 실험에서는 BCI 영어단어 학습 프로그램에 제안한 방법을 적용하여 두 피험자 파형을 비교했다. 실험에서는 두 피험자의 파형 유사 정도를 상관계수로 측정했다. 명상 값은 제안한 방법을 적용할 때 약 13%가 증가되었고 집중 값은 약 8%정도 증가되었다. 제안한 정규화 방법은 뇌파에 반영된 개인의 특성을 줄여서 제어에 적합한 파형으로 변환하기 때문에 게임과 같은 응용프로그램에 적합하다. In the beginning brain waves were used for monkeys to control robot arm with neural activity. In recent years there are research that measured brain waves are used for the control of programs which monitor the progression of dementia or enhance of attention in children diagnosed with Attention Deficit Hyperactivity Disorder (ADHD). Moreover, low-price devices that can be used as a game control interface have become available. One of the problems associated with control using brain waves is that the mean amplitude, mean wavelength, and mean vibrational frequency of the brain waves differ from individual to individual. This paper attempts to propose a method to normalize measured brain waves using normal distribution and calculate the waveforms that can be used in controlling games. For this, a framework in which brain waves are converted in seven stages has been suggested. In addition, the estimation process in each stage has been described. In an experiment the waveforms of two subjects have been compared using the proposed method in the BCI English word learning program. The level of similarity between two subjects' waveforms has been compared with correlation coefficient. When the proposed method was applied, both meditation and concentration increased by 13% and 8%, respectively. Because the proposed regularization method is converted into a waveform fit for control functions by reducing personal characteristics reflected in the brain waves, it is fitting for application programs such as games.

Future Trends of Blockchain and Crypto Currency: Challenges, Opportunities, and Solutions

성연식,박종혁 한국정보처리학회 2019 Journal of information processing systems Vol.15 No.3

The blockchain and crypto currency has become one of the most essential components of a communicationnetwork in the recent years. Through communication networking, we browse the internet, make VoIP phonecalls, have video conferences and check e-mails via computers. A lot of researches are being conducting toaddress the blockchain and crypto currency challenges in communication networking and providecorresponding solutions. In this paper, a diverse kind of novel research works in terms of mechanisms,techniques, architectures, and frameworks have been proposed to provide possible solutions against the existingchallenges in the communication networking. Such novel research works involve thermal load capacitytechniques, intelligent sensing mechanism, secure cloud computing system communication algorithm forwearable healthcare systems, sentiment analysis, optimized resources.

A Framework for Electroencephalogram Process at Real-Time using Brainwave

성연식,조경은,엄기현 한국멀티미디어학회 2011 멀티미디어학회논문지 Vol.14 No.9

Neurofeedback training using ElectroEncephaloGrams (EEGs) is commonly utilized in the treatment of Alzheimer’s disease, and Attention Deficit Hyperactivity Disorder (ADHD). Recently, BCI (Brain-computer Interface) contents have developed, not for the purpose of treatment, but for concentration improvement or brain relaxation training. However, as each user has different wave forms, it is hard to develop contents controlled by such different wave. Therefore, an EEG process that allows the ability to transform the variety of wave forms into one standard signal and use it without taking a user’s characteristic of EEG into account, is required. In this paper, a framework that can reduce users’ characteristics by normalizing and converting measured EEGs is proposed for contents. This framework also contains the process that controls different brainwave measuring devices. In experiment a handling process applying the proposed framework to the developed BCI contents is introduced.

A Framework for Processing Brain Waves Used in a Brain-computer Interface

성연식,조경은,엄기현 한국정보처리학회 2012 Journal of information processing systems Vol.8 No.2

Recently, methodologies for developing brain-computer interface (BCI) games using the BCI have been actively researched. The existing general framework for processing brain waves does not provide the functions required to develop BCI games. Thus, developing BCI games is difficult and requires a large amount of time. Effective BCI game development requires a BCI game framework. Therefore the BCI game framework should provide the functions to generate discrete values, events, and converted waves considering the difference between the brain waves of users and the BCIs of those. In this paper, BCI game frameworks for processing brain waves for BCI games are proposed. A variety of processes for converting brain waves to apply the measured brain waves to the games are also proposed. In an experiment the frameworks proposed were applied to a BCI game for visual perception training. Furthermore, it was verified that the time required for BCI game development was reduced when the framework proposed in the experiment was applied

A Method for Learning Macro-Actions for Virtual Characters Using Programming by Demonstration and Reinforcement Learning

성연식,조경은 한국정보처리학회 2012 Journal of information processing systems Vol.8 No.3

The decision-making by agents in games is commonly based on reinforcement learning. To improve the quality of agents, it is necessary to solve the problems of the time and state space that are required for learning. Such problems can be solved by Macro-Actions, which are defined and executed by a sequence of primitive actions. In this line of research, the learning time is reduced by cutting down the number of policy decisions by agents. Macro-Actions were originally defined as combinations of the same primitive actions. Based on studies that showed the generation of Macro-Actions by learning, Macro-Actions are now thought to consist of diverse kinds of primitive actions. However an enormous amount of learning time and state space are required to generate Macro-Actions. To resolve these issues, we can apply insights from studies on the learning of tasks through Programming by Demonstration (PbD) to generate Macro-Actions that reduce the learning time and state space. In this paper, we propose a method to define and execute Macro-Actions. Macro-Actions are learned from a human subject via PbD and a policy is learned by reinforcement learning. In an experiment, the proposed method was applied to a car simulation to verify the scalability of the proposed method. Data was collected from the driving control of a human subject, and then the Macro-Actions that are required for running a car were generated. Furthermore, the policy that is necessary for driving on a track was learned. The acquisition of Macro-Actions by PbD reduced the driving time by about 16% compared to the case in which Macro-Actions were directly defined by a human subject. In addition, the learning time was also reduced by a faster convergence of the optimum policies.

Arm Orientation Estimation Method with Multiple Devices for NUI/NUX

성연식,최룡,정영식 한국정보처리학회 2018 Journal of information processing systems Vol.14 No.4

Motion estimation is a key Natural User Interface/Natural User Experience (NUI/NUX) technology to utilizemotions as commands. HTC VIVE is an excellent device for estimating motions but only considers thepositions of hands, not the orientations of arms. Even if the positions of the hands are the same, the meaningof motions can differ according to the orientations of the arms. Therefore, when the positions of arms aremeasured and utilized, their orientations should be estimated as well. This paper proposes a method forestimating the arm orientations based on the Bayesian probability of the hand positions measured in advance. In experiments, the proposed method was used to measure the hand positions with HTC VIVE. The resultsshowed that the proposed method estimated orientations with an error rate of about 19%, but the possibilityof estimating the orientation of any body part without additional devices was demonstrated.

의사결정의 정확성을 고려한 Q-학습의 상태 공간을 줄이는 방법- 3D환경에 적용한 사례 중심으로

성연식,조경은,엄기현 동국대학교 영상문화콘텐츠연구원 2009 영상문화콘텐츠연구 Vol.0 No.-

A Technique to Solve the Q-learning's Curse of Dimensionality Considering Decision-Making's Accuracy - Focused on the Case Study in 3D Environment Sung, Yun-sick, Cho Kyung-eun, Um, Ky-hyun As Virtual reality which interacts with user extends its research fields, it started to integrate with artificial intelligence. If virtual human and animal is controlled using traditional artificial intelligence like FSM in Virtual reality, their actions are usually executed recursively or it's sometime vary difficult to define their actions when action's model is complex. Q-learning suits virtual environment, because agent could learn without the definition of environment model. But it needs to solve the curse of dimensionality, because state space is defined bigger as environment is more complex. The research which solves this problem reduces state space and calculates the Q-value which state is not defined. But it needs learning algorithm considering all states to decide an action accurately. In this paper, we propose two methods which can preserve decision's accuracy. First, we propose the method which preserves decision's accuracy using planner when Q-learning's state 244 space is reduced. Second, we introduce the way how to divide virtual environment and define state space of learning using numbers of small state spaces. In this way, we divided the space into two regions which could cut down the state space where learning results store by 53%. 가상환경에서 사용자와 상호 작용하는 가상현실 분야는 점점 넓어지면서인공지능 기술을 사용한다. 가상현실에서 사람이나 동물의 움직임을 제어할때 FSM과 같은 전통적인 인공지능을 사용하면 동일한 행동이 반복적으로 수행되거나 행동 모델이 복잡한 경우에는 정의하기가 어려운 단점이 있다. Q-학습은 환경 모델을 정의하지 않고 학습하기 때문에 가상환경에 적합하다. 하지만 환경이 복잡해지면 학습한 결과를 저장하는 상태 공간이 커지는 문제가 발생한다. 상태 공간 문제를 해결하기 위한 연구들은 학습에 적용하는 상태를줄여서 상태 공간을 줄인다. 그리고 발생한 오차는 계산하거나 추정하는 방법을 사용한다. 하지만 정확한 의사결정을 하기 위해서는 모든 상태를 고려하면서 학습하는 방법이 필요하다. 이 논문에서는 Q-학습의 상태 공간을 줄이면서의사결정의 정확성이 떨어지는 것을 보완할 수 있는 두 가지 방법을 제안한다. 첫 번째, 학습과 계획을 통합한 방법에서는 상태 공간을 줄일 때 발생하는의사결정의 정확성 문제를 계획기를 사용해서 보완하는 방법을 제안한다. 두번째, 가상환경의 공간을 분리해서 다수의 작은 상태공간으로 학습에 사용하는 상태공간을 정의하는 방법을 소개한다. 이 방법은 사례에서 공간을 두 개로 나누어 상태공간을 정의함으로써 학습 결과를 저장하는 상태공간을 약53%로 줄였다.

Programming by Demonstration을 이용한 가상 캐릭터의 행동 생성 기법

성연식(Sung, Yun-Sick),조경은(Cho, Kyung-Eun),엄기현(Um, Ky-Hyun) 한국게임학회 2011 한국게임학회 논문지 Vol.11 No.2

가상 환경에서 가상 캐릭터가 사람과 같이 자연스럽게 동작하기 위한 과정은 많은 노력을 필요로 한다. 가상 캐릭터는 행동을 기반으로 동작하기 때문에 자연스럽게 동작하기 위해서는 행동 정의가 가장 중요하다. 최근에는 사람이 조작한 가상 캐릭터의 행동을 기반으로 가상 캐릭터가 수행할 일련의 행동을 정의하는 Programming by Demonstration에 관한 연구가 활성화되고 있다. 하지만 이 방법들은 항상 동일한 길이의 연속적인 행동을 도출하거나 연속적인 행동을 정의할 때 사용하는 행동들을 제한하기 때문에 다양한 일련의 행동 생성하기가 어렵다. 자연스럽게 동작을 수행하기 위해서는 조작한 가상 캐릭터의 행동에서 최대한 다양한 일련의 행동들을 도출하고 도출한 일련의 행동들을 대표하는 최적의 연속적인 행동을 선택해야 한다. 그래서 여러 가지의 제약 사항을 줄이면서 일련의 행동을 생성하기 위한 연구들이 필요하다. 이 논문에서는 사람이 조작한 캐릭터의 행동을 수집하고 가상 캐릭터가 수행할 일련의 행동을 모두 도출해서 가상 캐릭터에게 적합한 일련의 행동을 선정하는 방법을 제안한다. 실험에서는 제안한 방법을 자동차 게임에 적용해서 생성되는 과정을 기술한다. 생성된 결과를 보면 사람이 수행한 일련의 행동을 길이나 분할하는 위치에 제약을 받지 않고 생성한 것을 볼 수 있었다. 결국, 제안한 방법은 더 많은 일련의 행동을 도출하고 가상 캐릭터에게 적합한 행동을 선정할 수 있음을 확인하였다. The substantial effort is required to make a virtual character smoothly move like a human being in the virtual environment. Since a virtual character acts on the basis of the actions, it is the most critical to define actions for smooth flow of action. It has been actively studied the programming by demonstration which defines series of actions to be performed by a virtual character based on the actions operated by a person. However, such approaches can't easily draw many sequential actions because they create sequential actions in the same length all the time or restrict the actions used to create such actions. For smooth flow of action, it is required to derive sequential actions as much as possible from the actions of a virtual character and to select the representative set of actions. Accordingly, it is necessary to study how to create sequential actions as reducing diverse limits. This study suggests the approach to select sequential actions suitable for a virtual character by collecting the actions of a character manipulated by a person and deriving a set of actions to be performed by a virtual character. The experiment describes the process to create the actions by applying the approach suggested in this study to the driving game. In accordance with the analysis results, it was found that a set of actions performed by a person was generated without being restricted by a length or a part to divide. Finally, we confirmed that the suggested method selects the best sequential actions, appropriate to virtual character, among more generated actions.

영향력 분포도를 이용한 Q-학습

성연식(Yunsick Sung),조경은(Kyungeun Cho) 한국멀티미디어학회 2006 멀티미디어학회논문지 Vol.9 No.5