|Topic:||Hash encoding for neural radiance fields|
|Supervisor:||Mgr. Jonáš Kulhánek|
|Announce as:||Diplomová práce, Bakalářská práce, Semestrální projekt|
|Description:||Novel view synthesis is a long-standing computer vision problem with a significant impact in other fields such as computer graphics and robotics. In this thesis, the student will work with the newest state-of-the-art neural rendering methods such as [1,2]. Previously, neural radiance field (NeRF) methods suffered long training times (~GPU days). However, recently, a novel approach was proposed  that learns in minutes and approaches the performance of traditional NeRF approaches. Unfortunately, the novel approach also introduces unwanted artifacts in the rendered 3D scene. This thesis aims to resolve the problems with the existing method by designing alternative hash strategies and storage mechanisms to resolve the issue.
Although this thesis can be challenging, if it goes as expected, the results can be presented or published in a prestigious conference or journal (e.g., CVPR, ICCV), which can benefit the student’s career greatly.
|Bibliography:||: Müller, T., Evans, A., Schied, C., & Keller, A. (2022). Instant Neural Graphics Primitives with a Multiresolution Hash Encoding. arXiv:2201. 05989.
: Barron, Jonathan T., et al. "Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields." arXiv preprint arXiv:2103.13415 (2021).