Vykreslování objemových vícemateriálových 3D výtisků v reálném čase
| Thesis title in Czech: | Vykreslování objemových vícemateriálových 3D výtisků v reálném čase |
|---|---|
| Thesis title in English: | Real-time rendering of volumetric multi-material 3D printouts |
| Academic year of topic announcement: | 2024/2025 |
| Thesis type: | diploma thesis |
| Thesis language: | |
| Department: | Department of Software and Computer Science Education (32-KSVI) |
| Supervisor: | Mgr. Tomáš Iser, Ph.D. |
| Author: | hidden - assigned and confirmed by the Study Dept. |
| Date of registration: | 28.04.2025 |
| Date of assignment: | 28.04.2025 |
| Confirmed by Study dept. on: | 28.04.2025 |
| Guidelines |
| The input for a full-colour 3D printer are volumetric data. As light scatters inside the volume, real-time rendering is impossible. In recent years, multiple techniques emerged for representing 3D appearance as a radiance field. They take images of the scene from different angles as input, and optimize their internal representation to match the images. Examples include Neural Radiance Fields (NeRF) and Gaussian splatting.
The goal of the thesis is to use such a technique to build a representation of a 3D printout that is only pre-computed once (that can be slow), but then visualized interactively from various angles. To build the representation, we will sample the 3D printout by casting individual rays from around it. This has a potential to make the rendering faster and achieve real-time previews close to the actual predictively-rendered ground truth. The student will: 1. Explore modern techniques that are used for photogrammetric reconstructions, such as NeRF and Gaussian splatting. 2. Understand how these algorithms work and use the one(s) that is/are the most suitable for the use case of multi-material volumetric 3D printing. 3. Propose an efficient method to sample the volumetric 3D printing data. 4. Implement an application to render the 3D printout in real time. 5. Measure runtimes and create benchmarks. |
| References |
| Scattering-aware texture reproduction for 3D printing
O Elek, D Sumin, R Zhang, T Weyrich, K Myszkowski, B Bickel, A Wilkie, ... ACM Transactions on Graphics 36 (6), 2017 Geometry-aware scattering compensation for 3D printing D Sumin, T Rittig, V Babaei, T Nindel, A Wilkie, P Didyk, B Bickel, ... ACM Transactions on Graphics 38 (4), 2019 A gradient-based framework for 3D print appearance optimization TK Nindel, T Iser, T Rittig, A Wilkie, J Křivánek ACM Transactions on Graphics (TOG) 40 (4), 1-15, 2021 NeRF: Representing scenes as neural radiance fields for view synthesis B Mildenhall, PP Srinivasan, M Tancik, JT Barron, R Ramamoorthi, R Ng arXiv preprint arXiv:2003.08934, 2020 3D Gaussian Splatting for Real-Time Radiance Field Rendering B Kerbl, G Kopanas, T Leimkühler, G Drettakis ACM Transactions on Graphics 42 (4), 2023 |