• Korean Institute of Interior Design Journal
• /
• v.26 no.6
• /
• pp.81-96
• /
• 2017

This study aims at establishing a basic theory for the combination of architecture and movies by comparing the long-take technique of movies and the continuity of space, one of space composition principles, which is important in digital architecture based on Jacques Lacan's visual-art theory and finding common features and differences of them. The following is a summary of the conclusions. First, analyzing the long-take technique on the basis of Lacan's visual-art theory found that the subject of representation is scenes of movies and that staring shows features of narrative. Second, the long-take technique can be thought as a cinematic technique which tries to realize the real order beyond the symbolic order in real life through the process of continuous replication of replication of replication of a scene in one shot. Third, in contemporary architecture, which is compared to the long-take technique in the past, the inclined space of opened gaze is similar to the method which tries to realize architectural space of the reality which belongs to the symbolic order close to the real order which belong to significant in human unconsciousness. Fourth, the freeform continuous space of closed gaze, which can be compared to contemporary long take combined with computer graphic technology, has more difficulty in realizing the real order than the long-take technique in the past and inclined, continuous space as the feature which belongs to $signifi{\acute{e}}$ in human consciousness has been strengthened through the circulation which repeats and expands along an observer's movement. Fifth, when the contemporary long-take technique and freeform continuous space expand gaze which opens from the inside to the outside, it is considered that the space which is closer to the real order than the classic long-take technique and inclined continuous space can be created.

• Journal of Broadcast Engineering
• /
• v.24 no.2
• /
• pp.273-280
• /
• 2019

360 video is attracting attention as immersive media with the spread of VR applications, and MPEG-I (Immersive) Visual group is actively working on standardization to support immersive media experiences with up to six degree of freedom (6DoF). In virtual space of omnidirectional 6DoF, which is defined as a case of degree of freedom providing 6DoF in a restricted area, looking at the scene at any viewpoint of any position in the space requires rendering the view by synthesizing additional viewpoints called virtual omnidirectional viewpoints. This paper presents the performance results on view synthesis and their analysis, which have been done as exploration experiments (EEs) of omnidirectional 6DoF in MPEG-I. In other words, experiment results on view synthesis in various aspects of synthesis conditions such as the distances between input views and virtual view to be synthesized and the number of input views to be selected from the given set of 360 videos providing omnidirectional 6DoF are presented.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.2
• /
• pp.79-86
• /
• 2024

In this paper, a novel visual tracking method which can utilize the context of object regions is presented. Conventional methods have the inherent problem of treating all candidate regions independently, where the tracker could not successfully discriminate regions with similar appearances. This was due to lack of contextual modeling in a given scene, where all candidate object regions should be taken into consideration when choosing a single region. The goal of the proposed method is to encourage feature exchange between candidate regions to improve the discriminability between similar regions. It improves upon conventional methods that only consider a single region, and is implemented by employing the MLP-Mixer model for enhanced feature exchange between regions. By implementing channel-wise, inter-region interaction operation between candidate features, contextual information of regions can be embedded into the individual feature representations. To evaluate the performance of the proposed tracker, the large-scale LaSOT dataset is used, and the experimental results show a competitive AUC performance of 0.560 while running at a real-time speed of 65 fps.

• Journal of Computing Science and Engineering
• /
• v.2 no.1
• /
• pp.98-108
• /
• 2008

Terrain reconstruction from images is an ill-posed, yet commonly desired Structure from Motion task when compositing visual effects into live-action photography. These surfaces are required for choreography of a scene, casting physically accurate shadows of CG elements, and occlusions. We present a novel framework for generating the geometry of landscapes from extremely noisy point cloud datasets obtained via limited resolution techniques, particularly optical flow based vision algorithms applied to live-action video plates. Our contribution is a new statistical approach to remove erroneous tracks ('outliers') by employing a unique combination of well established techniques-including Gaussian Mixture Models (GMMs) for robust parameter estimation and Radial Basis Functions (REFs) for scattered data interpolation-to exploit the natural constraints of this problem. Our algorithm offsets the tremendously laborious task of modeling these landscapes by hand, automatically generating a visually consistent, camera position dependent, thin-shell surface mesh within seconds for a typical tracking shot.

• Korean Journal of Computational Design and Engineering
• /
• v.15 no.3
• /
• pp.167-177
• /
• 2010

In recent years, the use of high-resolution tiled display system which does not have restrictions on the size of the screen and implements various layout of tile is increasing in order to evaluate the digital mock-up in physical scale or explore large engineering data set in detail. In this study, we developed multi-channel distributed visualization system which provides a virtual reality-based visual contents using 3D open-source graphics engine. Efficient data structures and exchange methods were proposed as a scene synchronization technology in PC cluster environments. DLP-Cube based tiled visualization system which provides $5{\times}2$ layout of display wall was developed and we validated our approach using this system. In addition, we introduced integrated control program that administrates PC cluster environment in remote and controls the layout of display channels.

• Journal of Korea Multimedia Society
• /
• v.18 no.8
• /
• pp.988-995
• /
• 2015

In 2013, Academy Award for Best Animated Short Film was granted to Walt Disney's short animation, (2012). With various aspects of its excellence, I focus on the very effective use of digital lightings and underscores for storytelling as its success factors. In this respect, this paper aims at analyzing the roles of the visual factors, especially tone, contrast, etc. created by lightings, and audio factors, especially underscores, in the film's story development. I find that can be characterized by the well-built story structure with distinct three acts. The main stream of the story is expressed with the overall mood that is created by the fine adjustments of brightness of the main light, and contrast. And the direction and the intensity of the lighting successfully describe the emotions of the characters in each scene. In addition, I find that properly chosen and positioned underscores make the development of the story more dynamic and more harmonized.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.35-36
• /
• 2016

본 논문에서는 기존 Bag-of-Visual words (BoW) 접근법에서 반영하지 못한 이미지의 공간 정보를 활용하기 위해서 Spatial Pyramid Matching (SPM) 기법을 Latent Dirichlet Allocation (LDA) 모델에 결합하여 이미지를 분류하는 모델을 제안한다. BoW 접근법은 이미지 패치를 시각적 단어로 변환하여 시각적 단어의 분포로 이미지를 표현하는 기법이며, 기존의 방식이 이미지 패치의 위치정보를 활용하지 못하는 점을 극복하기 위하여 SPM 기법을 도입하는 연구가 진행되어 왔다. 또한 이미지 패치를 정확하게 표현하기 위해서 벡터 양자화 대신 희소 부호화 기법을 이용하여 이미지 패치를 시각적 단어로 변환하였다. 제안하는 모델은 BoW 접근법을 기반으로 위치정보를 활용하는 SPM 을 LDA 모델에 적용하여 시각적 단어의 토픽을 추론함과 동시에 multi-class SVM 분류기를 이용하여 이미지를 분류한다. UIUC 스포츠 데이터를 이용하여 제안하는 모델의 분류 성능을 검증하였다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.11a
• /
• pp.290-294
• /
• 2001

비디오 데이터를 인덱싱 하기 위해서는 우선적으로 장면 전환을 검출하여 키 프레임을 추출하고 추출된 키 프레임을 바탕으로 인덱싱 작업을 수행한다. 본 논문에서는 장면 전환을 검출하기 위하여 컬러 히스토그램과 $\chi$$^2$히스토그램을 합성한 방법을 이용하여 키 프레임을 추출하고, 축구 비디오가 갖는 특성을 이용하여 샷 사이의 흐름을 파악하여 시각 정보를 분석하며, 이를 바탕으로 축구 비디오를 다양한 방법으로 인덱싱하는 방법을 제시한다.

• Advances in robotics research
• /
• v.1 no.2
• /
• pp.127-140
• /
• 2014

This work presents an IR-based system for parking assistance and obstacle detection in the automotive field that employs the Microsoft Kinect camera for fast 3D point cloud reconstruction. In contrast to previous research that attempts to explicitly identify obstacles, the proposed system aims to detect "reachable regions" of the environment, i.e., those regions where the vehicle can drive to from its current position. A user-friendly 2D traversability grid of cells is generated and used as a visual aid for parking assistance. Given a raw 3D point cloud, first each point is mapped into individual cells, then, the elevation information is used within a graph-based algorithm to label a given cell as traversable or non-traversable. Following this rationale, positive and negative obstacles, as well as unknown regions can be implicitly detected. Additionally, no flat-world assumption is required. Experimental results, obtained from the system in typical parking scenarios, are presented showing its effectiveness for scene interpretation and detection of several types of obstacle.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.4
• /
• pp.353-364
• /
• 2009

In this paper, a SLAM (Simultaneous Localization and Mapping) method using a single camera and planar fiducial markers is proposed. Fiducial markers are planar patterns that are mounted on the ceiling or wall. Each fiducial marker has a unique hi-tonal identification pattern with square outlines. It can be printed on paper to reduce cost or it can be painted using retro-reflective paint in order to make invisible and prevent undesirable visual effects. Existing localization methods using artificial landmarks have the disadvantage that landmark locations must be known a priori. In contrast, the proposed method can build a map and estimate robot location even if landmark locations are not known a priori. Hence, it reduces installation time and setup cost. The proposed method works good even when only one fiducial marker is seen at a scene. We perform computer simulation to evaluate proposed method.