AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

The Surprising Effectiveness of Noise Pretraining for Implicit Neural Representations

Kushal Vyas, Alper Kayabasi, Daniel Kim, Vishwanath Saragadam, Ashok Veeraraghavan, Guha Balakrishnan

CVPR 2026 · Architectures and efficiency › RNNs and normalization · approach: Sparse dictionary representations · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
pretrain INR on unstructured Uniform noise instead of real images (SNP: Uniform vs STRAINER) (ablation)STRAINER-style SIREN encoder · CelebA-HQ image fittingPSNR (dB): 57.8 -> 85.7Improved
pretrain on Gaussian noise instead of real images (ablation)STRAINER-style SIREN encoder · CelebA-HQ image fittingPSNR (dB): 57.8 -> 80.0Improved
pretrain on 1/f spectrum noise instead of real face images (ablation)STRAINER-style SIREN encoder · CelebA-HQ image fittingPSNR (dB): 57.8 -> 56.4No clear effect
pretrain on noise instead of real images, evaluated on OASIS-MRI (ablation)STRAINER-style SIREN encoder · OASIS-MRI image fittingPSNR (dB): STRAINER 62.80 -> SNP: Uniform 79.3Improved
use Dead Leaves noise pretraining instead of spectrum noise (ablation)SNP · CelebA-HQ image fittingPSNR (dB): 56.4 -> 55.0Hurt
add SNP pretraining to SIREN INR (any pretraining on noise)SIREN · CelebA-HQ image fittingPSNR (dB): 44.9 -> 56.4 (Spectrum)Improved
use unstructured noise (Uniform, Gaussian) pretraining instead of SIREN random init (ablation)INR denoising · CelebA-HQ denoisingPSNR (dB): SIREN 28.3 -> Uniform 23.0, Gaussian 23.8Hurt
use spectrum noise pretraining instead of SIREN random init for denoisingINR denoising · CelebA-HQ denoisingearly-stopping iteration: 139 -> 78Improved
add SNP: Spectrum initialization to ResFieldsResFields · Pexels video fittingmean PSNR (dB): 29.5 -> 31.1Improved
add SNP: Spectrum initialization to ResFieldsResFields · Pexels video denoisingmean PSNR (dB): 27.0 -> 28.0Improved
use Uniform instead of Spectrum noise for video (ablation)ResFields + SNP · Pexels video fittingmean PSNR (dB): 31.1 -> 30.9Hurt
choose Uniform noise over Spectrum noise pretraining (ablation)SNP · fitting vs denoising tradeoff, CelebA-HQfitting PSNR vs denoising PSNR: fitting 56.4 -> 85.7; denoising 27.6 -> 23.0Mixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy