Research map / 3D vision / Face and shape
Recurrent Sinusoidal INRs for Efficient High-Fidelity Representation
Hyunmin Cho, Jaejun Yoo, Kyong Hwan Jin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| increase recurrent unrolling steps from feed-forward to R=5 (ablation) | Recurrent sinusoidal INR · Set5-style fitting at step 500 | PSNR (dB): 39.724 -> 63.378 | Improved |
| add weight-tied recurrence to SIREN (ablation) | SIREN · Capacity-limited fitting, ~600K | PSNR (dB): 48.64 -> 78.35 | Improved |
| add weight-tied recurrence to FINER (ablation) | FINER · Capacity-limited fitting, ~600K | PSNR (dB): 48.36 -> 85.62 at ~600K; 38.01 -> 34.21 at ~100K | Mixed |
| add weight-tied recurrence to Gauss (ablation) | Gauss · Capacity-limited fitting, ~600K | PSNR (dB): 60.31 -> 12.02 | Hurt |
| add weight-tied recurrence to PEMLP (ablation) | PEMLP · Capacity-limited fitting, ~600K | PSNR (dB): 34.20 -> 31.04 | No clear effect |
| add recurrent bias term in recurrent layer (ablation) | Recurrent sinusoidal INR · fitting with R>=2 | PSNR (dB) | Hurt |
| add binarized supervision on top of recurrence (ablation) | Recurrent FINER-style INR · Kodak24 | PSNR (dB): 64.19 -> inf; bit error 49.70K -> 0 | Improved |
| add recurrence to feed-forward baseline (ablation) | FINER feed-forward · Kodak24 | PSNR (dB): 42.15 -> 64.19 | Improved |
| replace natural binary coding with Gray coding (ablation) | Ours · Kodak24 x2 SR | SR PSNR (dB): 23.45 -> 26.55 | Improved |
| replace natural binary coding with Gray coding (ablation) | RFF-F10 · Kodak24 x2 SR | SR PSNR (dB): 21.08 -> 23.98 | Improved |
| replace natural binary coding with Gray coding (ablation) | Ours · LLFF Flower NeRF | PSNR (dB): 21.14 -> 23.80 | Improved |
| replace natural binary coding with Gray coding (ablation) | Baseline NeRF · LLFF Flower NeRF | PSNR (dB): 20.79 -> 23.05 | Improved |
Closest papers by meaning
- A sampling theory perspective on activations for implicit neural representations (ICML 2024 · similarity 0.88)
- FINER: Flexible Spectral-bias Tuning in Implicit NEural Representation by Variable-periodic Activation Functions (CVPR 2024 · similarity 0.87)
- FreSh: Frequency Shifting for Accelerated Neural Representation Learning (ICLR 2025 · similarity 0.87)
- Tuning the Frequencies: Robust Training for Sinusoidal Neural Networks (CVPR 2025 · similarity 0.86)
- Spectral Alignment in Forward–Backward Representations via Temporal Abstraction (NeurIPS 2026 · similarity 0.86)
- Improved Implicit Neural Representation with Fourier Reparameterized Training (CVPR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
