Research map / Architectures and efficiency / RNNs and normalization
Spectral Gating via Damped Oscillations for Adaptive Implicit Neural Representations
Alex Costanzino, Pierluigi Zama Ramirez, Giuseppe Lisanti, Luigi Di Stefano
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 |
|---|---|---|---|
| use FDHO activation instead of SIREN | INR · Square Wave 500 Hz | PSNR (dB): 64.49 -> 126.83 | Improved |
| use FDHO activation instead of SIREN | INR · Chirp 1000 Hz | PSNR (dB): 61.05 -> 86.39 | Improved |
| use FDHO instead of FINER for peak PSNR | INR · Chirp 1000 Hz (peak PSNR) | PSNR (dB): 112.57 -> 94.15 | Hurt |
| use FDHO instead of Fourier Features for peak PSNR | INR · Chirp 250 Hz (peak PSNR) | PSNR (dB): 137.95 -> 120.97 | Hurt |
| use FDHO instead of SIREN | INR · Audio Bach (final PSNR) | PSNR (dB): 52.02 -> 56.23 | Improved |
| use FDHO instead of SIREN for peak PSNR | INR · Audio Bach (peak PSNR) | PSNR (dB): 58.47 -> 56.32 | Hurt |
| use FDHO instead of FINER | INR · Image fitting Bikers (final PSNR) | PSNR (dB): 47.46 -> 57.06 | Improved |
| use FDHO instead of FINER | INR · Image fitting Tiles (final PSNR) | PSNR (dB): 54.09 -> 52.24 | Hurt |
| use FDHO instead of SIREN | INR · Image denoising (sigma=0.1) | PSNR (dB): 26.22 -> 30.81 | Improved |
| use FDHO instead of MFN | INR · CT reconstruction | PSNR (dB): 37.19 -> 39.97 | Improved |
| use FDHO instead of SIREN | INR · CT reconstruction | PSNR (dB): 29.81 -> 39.97 | Improved |
| use FDHO instead of SIREN | INR · Image inpainting Tiger (20% pixels) | PSNR (dB): 27.35 -> 26.86 | Hurt |
Closest papers by meaning
- A sampling theory perspective on activations for implicit neural representations (ICML 2024 · similarity 0.90)
- Tuning the Frequencies: Robust Training for Sinusoidal Neural Networks (CVPR 2025 · similarity 0.88)
- FreSh: Frequency Shifting for Accelerated Neural Representation Learning (ICLR 2025 · similarity 0.88)
- FINER: Flexible Spectral-bias Tuning in Implicit NEural Representation by Variable-periodic Activation Functions (CVPR 2024 · similarity 0.87)
- Inductive Gradient Adjustment for Spectral Bias in Implicit Neural Representations (ICML 2025 · similarity 0.86)
- Spectral Alignment in Forward–Backward Representations via Temporal Abstraction (NeurIPS 2026 · similarity 0.86)
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.
