from HyperNeRF: A Higher-Dimensional Representation for Topologically Varying Neural Radiance Fields
- Wow, this technique is quite old, dating back to 1988.
- It seems like this approach treats something that changes (e.g., a 2D shape) as slices of a higher-dimensional object (e.g., a 3D object).
- Check out this video for a visual demonstration: https://hypernerf.github.io/static/figures/level_set/interpolate2.mp4
- By capturing moving scenes in a higher dimension, they can be represented as smooth transformations instead of discontinuous ones.
- Ah, I see. That’s really interesting! (blu3mo)