AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

SHACIRA: Scalable HAsh-grid Compression for Implicit Neural Representations

Sharath Girish, Abhinav Shrivastava, Kamal Gupta

ICCV 2023 · Architectures and efficiency › RNNs and normalization · approach: Dataset condensation and flow estimation · 33 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
reparameterize feature grids with quantized latents and entropy regularization on large imagesInstant-NGP · Cosmic-Cliffs imageBPP: 0.63 -> 0.11 BPP at 38.78 -> 38.79 PSNRImproved
compress Instant-NGP feature grid with SHACIRA on RTMV NeRF scenesInstant-NGP · RTMV 10 brick scenesStorage: 48.9MB -> 1.03MB; PSNR 31.88 -> 32.14Improved
add entropy regularization on latents (ablation)SHACIRA · UVG-F Jockey imagePSNR-BPP trade-offImproved
increase entropy regularization weight to 4.0e-4 (ablation)SHACIRA · UVG-F Jockey imagePSNR-BPP trade-offHurt
add annealing of discrete latents instead of STE alone (ablation)SHACIRA · UVG-F Jockey imagePSNR-BPP trade-offImproved
use feature-grid SHACIRA instead of MLP-based positional INRPositional INR · UVG-FPSNR: 33.17 -> 37.74Improved
use SHACIRA instead of NeRV for videoUVG video · UVG videoPSNR: 35.54 -> 35.01 PSNR; 0.66 -> 0.21 BPPMixed
use SHACIRA feature grid vs MLP-based INR for convergence speedKodak image · Kodakencoding time: 600s vs 53s at similar PSNR and BPPImproved
use SHACIRA at very high resolution (Pearl, Tokyo)Instant-NGP · Pearl and TokyoPSNR: 29.62 -> 29.44; 31.87 -> 31.62Mixed
use SHACIRA compared with VQADVQAD · RTMV NeRFStorage: 0.55MB -> 0.43MB; PSNR 31.45 -> 31.46Improved

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