A5004169490- Advanced Vision and Imaging
- Computer Graphics and Visualization Techniques
- Spectroscopy and Chemometric Analyses
- Advanced Image and Video Retrieval Techniques
- Advanced Image Processing Techniques
- Computational Geometry and Mesh Generation
- 3D Surveying and Cultural Heritage
- Satellite Image Processing and Photogrammetry
Sungkyunkwan University
2020-2021
Abstract This paper presents a scalable online occlusion culling algorithm, which significantly improves the previous raster using object‐level bounding volume hierarchy. Given occluders found with temporal coherence, we find and rasterize coarse groups of potential occludees in Within rasterized bounds, per‐pixel ray casting tests fine‐grained visibilities every individual occludees. We further propose acceleration techniques including read‐back counters for tightly‐packed multidrawing...
This article presents a real-time bokeh rendering technique that splats pre-computed sprites but takes dynamic visibilities and intrinsic appearances into account at runtime. To attain alias-free looks without excessive sampling on lens, the of strong highlights are densely sampled using rasterization, while regular objects sparsely conventional defocus-blur rendering. The appearance is dynamically transformed from precomputed look-up table, which encodes radial aberrations against image...
희소옥트리 기반의 복셀콘추적 렌더링 알고리즘은 전역 조명에 필요한 정보를 볼륨에 저장하고, 복셀콘추적을 통해 간접 확산과 반사 효과를 효율적으로 계산할 수 있다. 그러나 거울 정반사와 같이 고해상도가 경우에 복셀콘추적은 현격한 메모리와 연산을 필요로 한다. 본 연구는 희소옥트리의 잎노드에서 프리미티브가 밀집되지 않은 경우는 프리미티브 목록을 사용하고, 밀집된 추가 복셀화를 수행하는 하이브리드 옥트리와 콘추적과 광선추적을 노드 형태에 따라 적응적으로 추적 기법을 제안한다. 기법은 고품질 간접조명을 복셀콘추적보다 빠르게 실시간에 수행할