• Journal of Korea Game Society
• /
• v.1 no.1
• /
• pp.55-60
• /
• 2001

With rapid development of graphic hardware, researches on Virtual Reality and 3D Games have received more attention than before. For more realistic 3D graphic scene, objects were to be presented with lots of polygons and the number of objects shown in a scene was remarkably increased. Therefore, for effective visualization of large polygon models like this, view-frustum culling method, that visualizes only objects shown in the screen, has been widely used. In general, the bounding boxes that include objects are generated firstly, and the boxes are intersected with view-frustum to check whether object is in the visible area or not. Recently, an algorithm that can check in-out test of objects using OpenGL's selection mode, which is originally used to select the objects in the screen, is suggested. This algorithm is fast because it can use hardware acceleration. In this study, by implementing and applying this algorithm to large polygon models, we showed the efficiency of OpenGL assisted View-Frustum Culling algorithm. If this algorithm is applied to 3D games that have to process more complicated characters and landscapes, performance improvement can be expected.

• The KIPS Transactions:PartA
• /
• v.9A no.4
• /
• pp.581-588
• /
• 2002

In implementing interactive multimedia systems, real-time visualization plays an important role. This paper presents efficient data structures and algorithms for real-time terrain navigation. Terrain data set is usually too huge to display as is. Therefore LOD (levels of detail) methods and view frustum culling are essential tools. This paper describes in detail compact hierarchical data structures, fast view frustum culling, and efficient LOD construction/rendering algorithms. Unlike previous works, we use a precise screen-space error metric for vertex removal and a strict error threshold allowing sub-pixel -sized errors only. Nevertheless, we can achieve 22 fps on average in a PC platform. The methods presented in this paper also satisfy almost all of the requirements for interactive real-time terrain Visualization.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2005.05a
• /
• pp.835-841
• /
• 2005

실시간 가시화/시각화(Visualization)을 위한 지형 가시화 분야에서의 렌더링 속도 향상은 사용자의 현실감에 있어 중요한 역할을 한다. 본 연구는 Height Field 데이터를 사용한 블록 단위 지형 렌더링 방법에서 이전 프레임 정보를 사용하여 지형 가시화 속도를 향상시키는 방법을 제안한다. Height Field로 표현된 지형을 실시간에 효과적으로 렌더링 할 수 있는 방법으로CLOD(Continuous leveI Of Detail)가 있으며 대표적인 방법으로 Multiresolution TIN(Triangle Irregular Network)과 Regular Grid Method에 기반한 방법들이 있다. 대개의 경우, 지형 데이터는 매우 방대한 크기를 가지고 있어서 실시간으로 렌더링 하는 것이 불가능할 경우가 많다. 따라서 실시간 지형렌더링 에서는 LOD(Level of Detail)관리와 뷰 프러스텀 컬링이 실시간 렌더링 속도 향상을 위한 핵심 사항이 된다. 본 연구에서는 PC 기반에서 효과적으로 지형을 표현하기 위해 기존 Regular Grid Method에 기반한 방법의 가시영역 추출(View Frustum Culling)을 수정하여 실시간 지형 렌더링의 가시 영역 추출(View Frustum Culling)시 기존의 쿼드트리 방식과 함께 이전 프레임에서 블록 단위로 저장된 컬링 정보를 혼용하여 속도를 향상시키는 방법을 상세히 기술 한다.

• Journal of Korea Game Society
• /
• v.7 no.1
• /
• pp.43-52
• /
• 2007

This article describes a new contribution culling method for fast rendering of complex huge urban scenes. A view frustum culling technique is used for fast rendering of complex scenes. To support the levels-of-detail, we subdivide the image regions and construct a weighted quadtree. Only visible objects at the current camera position contributes the current quadtree and the weight is assigned to each object in the quadtree. The weight is proportional to the image area of the projected object, so large buildings in the far distance are less likely to be culled out than small buildings in the near distance. The rendering time is nearly constant not depending on the number of visible objects. The proposed method has applied to a new metropolitan region which is currently under development. Experimental results showed that the rendering quality of the proposed method is barely distinguishable from the rendering quality of the original method, while the proposed method reduces the number of polygons by about 9%. Experimental results showed that the proposed rendering method is appropriate for real-time rendering applications of complex huge scenes.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.2
• /
• pp.146-155
• /
• 2008

To improve the performance of tiled display system, two parallel tiled display algorithms are proposed by combining the sort-first and the sort-last sorting methods. In the proposed algorithms, the view frustum culling is employed along with the OpenGL display list for the sort-first sorting, and the pre-detection sort-last sparse sorting method is used for sort-last sorting. Through the benchmarking tests, the performances of two proposed algorithms are investigated. Based on the observations, it is suggested how to select an optimal algorithm among the two proposed parallel tiled display algorithms for the given visualization model.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.10
• /
• pp.6917-6924
• /
• 2015

The universal nature of humans is formed by the view angle and the visibility range. However, the majority of theories on spatial structure analysis based on the visual perception do neither reflect the view angle nor consider only the flat view angle. Some theories that reflect them is a theory where the part included in the view angle and the part excluded in the view angle have been separated in a dichotomous way, excluding the universal characteristics of humans. This study applied an observing probability to a 3-D visibility analysis theory by conducting a eye-tracking experiment, empirically determining the limits of the field of view, and deriving the observing probability by view angle. In addition, it attempted to identify the probability by manufacturing an application of spacial, visual perception analysis and applying the concept of multiple frustum culling. For the characteristics of observation, the data were measured and collected regarding the walking course for 3 minutes for an optional space, aimed for 33 people as subjects. Subsequently, the data were prepared by analyzing the observation fixation frequency probability.

• Journal of Korea Game Society
• /
• v.8 no.3
• /
• pp.61-68
• /
• 2008

Quadtree-based terrain visualization methods have been used in a lot of applications. However, because most procedures are performed on the CPU, the rendering speed is slow in comparison to methods using GPU. In this paper, we present a quadtree-based terrain visualization method working on the GPU with specially designed data structure, error-texture and LOD-texture, and block-based acceleration method. In preprocessing step, we calculate errors in world space and store them to error-texture. In rendering step, we examine projected errors of error-texture and choose the detail level, then store the projected errors to LOD-texture. View frustum culling is performed as block unit using the values of error-texture and LOD-texture. This method reduces CPU load and performs time consuming jobs such as LOD selection and view frustum culling.

• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10b
• /
• pp.595-597
• /
• 1999

기존의 LOD 제어방법들은 랜더링속도를 성공적으로 증가시켜왔으나 오버헤드가 크다는 단점을 갖고 있다. 이러한 오버헤드는 각 vertex마다 view-frustum clipping, back-face culling, 스크린 공간 기하학적 오차계산과 같은 view-dependent refinement criteria를 측정하고, 메쉬의 LOD를 바꾸기 위해서 edge collapse/vertex split를 수행하기 때문이다. 제안하는 방법은 메쉬를 여러 개의 region들로 나누고 vertex가 아닌 region에 대해 view-dependent refinement criteria를 측정하므로 오버헤드가 훨씬 작다. 또한 각 region 들의 LOD가 바뀔 때 미리 만들어 둔 LOD 버전들중에서 하나를 선택하기만 하면 되므로, edge collapse/vertex split을 수행하는 오버헤드는 없다. 실험적으로 제안하는 LOD 제어방법은 기존의 방법들보다 작은 메모리를 사용하고 LOD 제어 오버헤드도 적으며, LOD 제어를 하지 않은 경우보다 2배-5배의 랜더링 속도향상을 얻었다.

• Journal of Advanced Navigation Technology
• /
• v.15 no.4
• /
• pp.616-625
• /
• 2011

There are two representative geometry rendering methods. One is Geometry Clipmaps, another is ROAM 2.0. We propose an extended Geometry Clipmaps algorithm which does not focus on CPU operation but the GPU for faster and wider visibility area. The extended algorithm presents mesh configuration method of each level by LOD, how to configurate Mesh network between levels, mesh block method for rendering optimization using VFC, and image mapping method to get high resolution up to 1 m.

• Journal of the Korea Computer Graphics Society
• /
• v.15 no.3
• /
• pp.27-33
• /
• 2009

In terrain visualization, the quadtree is the most frequently used data structure for progressive mesh generation. The quadtree provides an efficient level-of-detail selection and view frustum culling. However, most applications using quadtrees are performed by the CPU, since the hierarchical data structure cannot be manipulated in a programmable rendering pipeline. For this reason, quadtree-based methods show lower performance and higher dependancy of CPU in comparison to GPU-based methods. We present a quadtree-based terrain-rendering method for GPU execution that uses vertex multiplication. It offers higher performance than previous CPU-based quadtree methods, without loss of image quality.