• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.04a
• /
• pp.848-850
• /
• 2015

본 논문에서는 픽셀 그레디언트의 방향성 정보를 이용하여 복원하는 예제기반 인페인팅 방법을 제안한다. 제안하는 방법에서는 영상 내에서 픽셀 그레디언트의 방향성을 검출하여 강도가 높은 그레디언트를 가지는 픽셀들을 먼저 복원하는 특징을 가진다. 지역적인 그래디언트의 특성을 이용하여 복원 영역의 연결성을 높이고 에러의 파급 효과를 줄임으로써, 기존의 예제기반 인페인팅 방법보다 개선된 결과를 보여준다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.1339-1342
• /
• 2022

예술적 스타일 전이는 예술 작품이 지닌 특징을 다른 이미지에 적용하는 이미지 처리의 오랜 화두 중 하나로, 최근에는 StyleGAN 과 같이 미리 학습된 GAN(생성적 적대 신경망)을 통해 제한된 데이터로도 고해상도의 예술적 초상화를 생성하도록 학습하는 연구가 다양한 방면에서 성과를 내고 있다. 본 논문에서는 2 가지 경로의 StyleGAN과 Facial Destylization 을 통해 고해상도의 예시 기반 스타일 전이를 달성한 DualStyleGAN 연구에 대해 소개하고, 기존 연구에서 사용된 Facial Destylization 방법이 지닌 한계점을 분석한 뒤, 이를 개선한 새로운 방법, Re-Destyle을 제안한다. 새로운 Re-Destyle 방법으로 Facial Destylization 을 적용할 경우 학습 시간을 기존 연구의 방법보다 20 배 이상 개선할 수 있으며 그 결과 1000 개 이하의 적은 데이터와 1~2 시간의 추가 학습만으로도 원하는 타겟 초상화 스타일에 대해 1024×1024 수준의 고해상도의 예시 기반 초상화 스타일 전이 및 이미지 생성 모델을 학습할 수 있다.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.6
• /
• pp.395-402
• /
• 2013

We propose an accelerometer-based gesture recognition method for smartphone users. In our method, similarities between a new time series accelerometer data and each gesture exemplar are computed with DTW algorithm, and then the best matching gesture is determined based on k-NN algorithm. In order to investigate the performance of our method, we implemented a gesture recognition program working on an Android smartphone and a gesture-based teleoperating robot system. Through a set of user-mixed and user-independent experiments, we showed that the proposed method and implementation have high performance and scalability.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.44 no.1
• /
• pp.50-58
• /
• 2007

In this paper, we propose a query-based visual attention algorithm for effective object finding of a vision-based mobile robot. This algorithm is developed by extending conventional bottom-up visual attention algorithms. In our proposed algorithm various conspicuity maps are merged to make a saliency map, where weighting values are determined by query-dependent object properties. The saliency map is then used to find possible attentive location of queried object. To show the validities of our proposed algorithm, several objects are employed to compare performances of our proposed algorithm with those of conventional bottom-up approaches. Here, as one of exemplar query-dependent object property, color property is used.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.1
• /
• pp.183-200
• /
• 2020

This paper presents an efficient interactive framework for perspective scene completion and editing tasks, which are available largely in the real world but rarely studied in the field of image completion. Considering that it is quite hard to extract perspective information from a single image, this work starts from a friendly and portable interactive platform to obtain the basic perspective data. Then, in order to make this interface less sensitive, easier and more flexible, a perspective-rectification based correction mechanism is proposed to iteratively update the locations of the initial points selected by users. At last, a complanate mesh is generated by the geometry calculations from these corrected initial positions. This mesh must approximate the perspective direction and the structure topology as much as possible so that the filling process can be conducted under the constraint of the perspective effects of the original image. Our experiments show the results with good qualities and performances, and also demonstrate the validity of our approaches by various perspective scenes and images.

• Korean Journal of Computational Design and Engineering
• /
• v.18 no.1
• /
• pp.10-20
• /
• 2013

The Yoon Jeung residence is a well-known and prototypical aristocratic house example showing typical and interesting characteristics of the traditional houses in the middle region of Korea. When it comes to understanding the architectonic characteristics of a Korean traditional house, it was not easy to actually capture the compositional and/or constructional knowledge of the building even though it is a wooden building with many wooden members showing their compositional relations. Now with the help of the BIM tool, the Building Information Modeling tool, we could actually be able to compare and analyze each member and their compositional relations. In this paper we examine the unique traditional composition method used in extending the building's structural sections utilizing the bended-joint characteristics of traditional buildings with exemplar case of the Yoon Jeung residence. Thereby we examine those relationships among three major compositional parts namely the plan based spatial compositions, the upper wooden compositions and the roof forms so as to specify the building's typical characteristics with reasonably acceptable causes. The inner and outer block of the residence are handled with their bended-joints with more detailed knowledge of categorization by way of joint relationships among members.

• Journal of Korean Elementary Science Education
• /
• v.37 no.2
• /
• pp.173-187
• /
• 2018

The purpose of this study is to clarify the specific terms of epistemic and epistemological by reviewing the literature on epistemological understanding of science learning, examine the necessity of epistemic discourse analysis based on the view of social epistemology, and provide an exemplar of practical epistemology analysis for elementary children's science learning. The review was conducted in terms of meaning and terminology about epistemic or epistemological approach to science learning, epistemology of/for science, and methodologies for epistemic discourse analysis. As an alternative way of epistemic discourse analysis in science classroom I employed practical epistemology analysis (by Wickman), evidence-explanation continuum (by Duschl), and DREEC diagram (by Maeng et al.). The methods were administered to an elementary science class for the third grade where children observed sedimentary rocks. Through the outcomes of analysis I sought to understand the processes how children collected data by observation, identified evidence, and constructed explanations about rocks. During the process of practical epistemology analysis the cases of four categories, such as encounter, stand-fast, gap, and relation, were identified. The sequence of encounter, stand fast, gap, and relation showed how children observed sedimentary rocks and how they came to learn the difference among the rocks. The epistemic features of children's observation discourse, although different from scientists' discourses during their own practices, showed data-only conversation, evidence-driven conversation, or explanation inducing conversation. Thus I argue even elementary children are able to construct their own knowledge and their epistemic practices are productive.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.07a
• /
• pp.369-372
• /
• 2011

본 논문에서는 가상 시점 생성 시 발생하는 홀 영역의 메움 기법에 관하여 기하학적 제약조건을 이용한 예제 기반 영상 인페인팅 기술을 제안한다. 생성된 가상 시점이 실제 카메라들의 중간에 위치하는 경우, 이 때 발생하는 홀 영역은 시점 간 보상으로 메울 수 있다. 하지만, 가상 시점이 실제 카메라들의 Field-of-View를 벗어난 경우에는 홀 영역에 대한 정보가 없기 때문에 이를 메우는데 어려운 문제가 따른다. 이러한 문제를 해결하기 위해 제안한 방법은 예제 기반의 영상 인페인팅 기술을 이용하여 홀 영역을 메운다. 제안한 방법은 홀 영역의 기하학적 특성을 제약조건으로 우선순위(Priority)를 계산한다. 최적의 패치를 찾기 위해 컬러 정보뿐만 아니라 깊이 정보(Depth)를 활용한다. 또한, 전경 물체의 텍스쳐로 인하여 잘못된 패치가 선택되는 문제를 해결하기 위해 적응적으로 패치 크기를 변경한다. 실험 결과에서 제안한 방법은 일반적인 홀 메움 방식과 비교하여 객관적 평가와 주관적 평가 모두에서 시각적으로 우수한 결과를 보여주었다.

• Journal of Institute of Control, Robotics and Systems
• /
• v.19 no.12
• /
• pp.1145-1151
• /
• 2013

Kinematic modeling is a prerequisite for motion planning and the control of mobile robots. In this paper, we proposed a new method of kinematic modeling for a type of mobile robot based on differential motion transformation. The differential motion implies a small translation and rotation in three-dimensional space in a small time interval. Thus, transformation of the differential motion gives the velocity relationship, i.e., Jacobian between two coordinate frames. Since the theory of the differential motion transformation is well-developed, it is useful for the systematic velocity kinematic modeling of mobile robots. In order to show the validity for application of the differential motion transformation, we obtained velocity kinematic models for a type of exemplar mobile robot including spherical ballbots.

• Journal of IKEEE
• /
• v.20 no.1
• /
• pp.31-44
• /
• 2016

This paper presents a hole-filling algorithm based on extrapolating spatial-temporal background information used in view synthesis for free-viewpoint television. A new background codebook is constructed and updated in order to extract reliable temporal background information. In addition, an estimation of spatial local background values is conducted to discriminate an adaptive boundary between the background region and the foreground region as well as to update the information about the hole region. The holes then are filled by combining the spatial background information and the temporal background information. In addition, an exemplar-based inpainting technique is used to fill the rest of holes, in which a priority function using background-depth information is defined to determine the order in which the holes are filled. The experimental results demonstrated that the proposed algorithm outperformed the other comparative methods about average 0.3-0.6 dB, and that it synthesized satisfactory views regardless of video characteristics and type of hole region.