• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10c
• /
• pp.310-312
• /
• 2002

볼륨 렌더링은 물체의 겉면만이 아니라 내부에 있는 모든 3차원 데이터를 이용해서 렌더링 하는 기법이다. 따라서 기존의 폴리곤 렌더링에선 불가능했던 물체 내부에 대한 표현이 가능하기 때문에 과학, 의료 분야 등 물체 전체에 대한 데이터 처리가 필요한 곳에서 많이 쓰이고 있다. 하지만 이러한 볼륨데이터의 크기는 일반적으로 1024$^3$Bytes 이상이기 때문에 기존의 단일 그래픽 가속기로는 메모리 크기나 연산 능력면에서 처리하기에 한계가 있다. 따라서 본 논문에서는 이런 기가급 볼륨데이터를 처리하기 위한 병렬 볼륨 렌더링 구조를 제시하고, 전송된 부분 이미지 합성을 위한 블렌딩 순서를 결정하는 시점 추적 (point-tracking)기법과 네트웍에 의한 성능저하를 최소화 할 수 있는 ‘프레임간 유사성(frame-to-frame coherency)을 이용한 RGBA데이터 압축기법’을 제안한다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.01a
• /
• pp.349-351
• /
• 2022

본 논문에서는 계산양이 큰 볼륨 렌더링을 구현할 수 있는 파이썬 기반의 CUDA(Computed Unified Device Architecture) 커널(Kernel) 디자인에 대해서 소개한다. 최근에 파이썬은 인공지능뿐만 아니라 서버, 보안, GUI, 데이터 시각화, 빅 데이터 처리 등 다양한 분야에서 활용이 되고 있기 때문에 인터페이스만을 위한 언어라는 색을 탈피한지 오래이다. 본 논문에서는 대용량 병렬처리 기법인 NVIDIA의 CUDA를 이용하여 파이썬 환경에서 커널을 디자인하고, 계산양이 큰 볼륨 렌더링이 빠르게 계산되는 결과를 보여준다. 결과적으로 C언어 기반의 CUDA뿐만 아니라, 상대적으로 개발이 효율적인 파이썬 환경에서도 GPU(Graphic Processing Unit)기반 애플리케이션 개발이 가능하다는 것을 볼륨 렌더링을 통해 보여준다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.86-89
• /
• 2016

본 논문에서는 멀티미디어 콘텐츠의 명료도를 향상시켜 사용자가 주변 환경과 관계없이 안정적인 볼륨에서 오디오 청취를 할 수 있는 다이얼로그 명료도 향상 알고리즘을 제안한다. 최근 모바일 기기들의 발달로 다양한 환경에서 영화, TV 또는 동영상 등의 멀티미디어 콘텐츠를 즐기는 일이 늘어나고 있다. 이러한 경우 시청자는 주변 환경에 따라 영상의 오디오 볼륨을 조절하게 되는데 주변 소음에 비하여 과하게 증폭된 볼륨은 주변에 피해를 끼치거나 고막에 손상을 일으킬 수 있으며 반대로 주변에 비해 너무 작은 오디오 볼륨은 시청을 어렵게 한다는 단점이 있다. 본 논문에서는 수신단에서 멀티미디어 콘텐츠의 오디오 신호로부터 다이얼로그 성분을 검출하여 음성 명료도 향상 알고리즘을 적용시켜 동일한 볼륨에서도 음성의 명료도를 높이는 알고리즘을 제안한다. 본 알고리즘은 다이얼로그를 검출하여 단순히 증폭 시키는 기존 기술들과 달리 전체 에너지는 유지하면서 명료도에 중요한 영향을 미치는 주파수 대역에 에너지를 집중시키는 에너지 재분배 방식을 이용해 동일한 볼륨에서도 더 높은 음성 명료도를 기대할 수 있다. 컴퓨터 시뮬레이션을 통해 본 논문에서 제안한 알고리즘이 명료도에 중요한 영향을 미치는 주파수대역을 적절히 증폭시킴을 확인할 수 있었다.

• Journal of the Korea Computer Graphics Society
• /
• v.20 no.4
• /
• pp.1-7
• /
• 2014

The volume rendering has become an important technique in many applications along with hardware development. Understanding and perception of volume visualization benefit from visual cues which are available from shading. Better visual cues can be obtained from global illumination models but it's huge amount of computation and extra GPU memory need cause a lack of interactivity. In this paper, in order to improve visual cues on volume rendering, we propose an image space occlusion shading model which requires no additional resources.

• Proceedings of the Korea Society of Information Technology Applications Conference
• /
• 2002.11a
• /
• pp.350-362
• /
• 2002

A new architecture called the Storage Area Network(SAN) was developed in response to the requirements of high availability of data, scalable growth and system performance In order to use SAN more efficiently, must SAN operating systems support storage virtualization concepts that allow users to view physical storage devices attached to SAN as a large volume virtually. A logical volume manager Days a key role in storage virtualization it realizes the storage virtualization by mapping logical addresses to physical addresses. In this paper, we design and implement an efficient and flexible mapping method for logical volume manager. The mapping method in this paper supports a snapshot that preserves a volume image at certain time and on-line reorganization to allow users to add or remove storage devices to SAN even while the system is running.

• Journal of Korea Game Society
• /
• v.9 no.6
• /
• pp.191-196
• /
• 2009

Volume rendering is a method which extracts meaningful information from volume data and visualizes those information. In general, since the size of volume data gets larger, it is very important to devise acceleration methods for interactive rendering speed. Min-max octree is data structure for high-speed volume rendering, however, its creation time becomes long as the data size increases. In this paper, we propose acceleration method of min-max octree generation using CUDA. Firstly, we convert one-dimensional array from volume data using space filling curve. Then we make min-max octree structures from the sequential array and apply them to acceleration of volume ray casting.

• Journal of Korea Multimedia Society
• /
• v.8 no.9
• /
• pp.1239-1247
• /
• 2005

Basically, objects are discriminated by transfer functions in volume rendering . However, in some cases objects cannot be discriminated only with transfer functions. In these cases, objects are pre-segmented with other methods, and visualized based on the segmentation information. In this paper we present a way of assigning per-object transparency in visualization of segmented volumes. Semi-transparent rendering is used to effectively give context information about the observed object. Per-object transparency can be used as a very effective visualization tool especially when it is difficult to adjust transfer functions to make the object semi-transparent. We present several interpretations of the meaning of per-object transparency, and corresponding variations of the algorithm. We show that efficient implementations for interactive use are possible, by presenting an implementation using general graphics hardware.

• Journal of Biomedical Engineering Research
• /
• v.23 no.6
• /
• pp.491-498
• /
• 2002

In medical visualization fields, anisotropic volume data are more common than isotropic ones. In this paper, we propose an efficient rendering method for anisotropic volume data, which directly computes the intensity of intermediate samples by interpolating the intensity of two corresponding voxels on consecutive slices. Unlike conventional rendering method, it does not require a preprocessing step for generating intermediate slices or additional memory for storing them. In order to evaluate the validity and performance of our method, we applied the method to shear-warp rendering algorithm. Experimental results show that this method improves rendering speed without significantly sacrificing the image quality.

• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 2002.11a
• /
• pp.350-362
• /
• 2002

A new architecture called the Storage Area Network(SAN) was developed in response to the requirements of high availability of data scalable growth, and system performance. In order to use SAN more efficiently, most SAN operating systems support storage virtualization concepts that allow users to view physical storage devices attached to SAN as a large volume virtually. A logical volume manager plays a key role in storage virtualization. It realizes the storage virtualization by mapping logical addresses to physical addresses. In this paper, we design and implement an efficient and flexible mapping method for logical volume manager. The nin method in this paper supports a snapshot that preserves a volume image at certain time and on-line reorganization to allow users to add or remove storage devices to SAN even while the system is running.

• Journal of Korea Society of Industrial Information Systems
• /
• v.9 no.2
• /
• pp.23-31
• /
• 2004

This paper focuses on a boundary surface based ray casting. In general the boundary surface based ray casting is processed in two stages. The first stage finds boundary surfaces and stores them into buffers. The second stage calculates a distance from a viewpoint to the voxels of the interested area by projecting boundary surfaces on the view plane, and then starts to traverse a volume data space with the distance. Our approach differs from the general boundary surface based ray casting in processing the first stage of it. Contrast to the typical boundary surface based ray casting where all boundary surfaces of volume data are stored into buffers, they are projected on the planes aligned to the axis of volume data coordinates and these projected data are stored into 6 buffers. Such maneuver shortens time for ray casting, and reduces memory usage because it can be carried out independently from the amount of the volume data.