PREF: Phasorial Embedding Fields for Compact Neural Representations
About
We present an efficient frequency-based neural representation termed PREF: a shallow MLP augmented with a phasor volume that covers significant border spectra than previous Fourier feature mapping or Positional Encoding. At the core is our compact 3D phasor volume where frequencies distribute uniformly along a 2D plane and dilate along a 1D axis. To this end, we develop a tailored and efficient Fourier transform that combines both Fast Fourier transform and local interpolation to accelerate na\"ive Fourier mapping. We also introduce a Parsvel regularizer that stables frequency-based learning. In these ways, Our PREF reduces the costly MLP in the frequency-based representation, thereby significantly closing the efficiency gap between it and other hybrid representations, and improving its interpretability. Comprehensive experiments demonstrate that our PREF is able to capture high-frequency details while remaining compact and robust, including 2D image generalization, 3D signed distance function regression and 5D neural radiance field reconstruction.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Neural Radiance Field Reconstruction | NeRF Synthetic (test) | Time18 | 12 | |
| Image Painting | Text (test) | PSNR28.329 | 5 | |
| Image Painting | Natural (test) | PSNR24.113 | 5 | |
| SDF regression | Armadillo | IoU99.02 | 5 | |
| SDF regression | Gargoyle | IOU99.05 | 5 |