Lightplane: Highly-Scalable Components for Neural 3D Fields
This addresses a critical scalability problem for researchers and practitioners in 3D reconstruction and generation, offering incremental improvements to existing methods.
The paper tackles the memory-intensive bottleneck in 2D-3D mapping for neural 3D fields by proposing Lightplane Render and Splatter components, which significantly reduce memory usage and enable processing of more and higher-resolution images with small costs.
Contemporary 3D research, particularly in reconstruction and generation, heavily relies on 2D images for inputs or supervision. However, current designs for these 2D-3D mapping are memory-intensive, posing a significant bottleneck for existing methods and hindering new applications. In response, we propose a pair of highly scalable components for 3D neural fields: Lightplane Render and Splatter, which significantly reduce memory usage in 2D-3D mapping. These innovations enable the processing of vastly more and higher resolution images with small memory and computational costs. We demonstrate their utility in various applications, from benefiting single-scene optimization with image-level losses to realizing a versatile pipeline for dramatically scaling 3D reconstruction and generation. Code: \url{https://github.com/facebookresearch/lightplane}.