towards end-to-end neuromorphic voxel-based 3d object reconstruction without physical priors

Reported on 3 benchmarks across 3 tasks · 1 paper · 3 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.