AI Paper Atlas

Research map / Image generation / Image restoration

MB-RACS: Measurement-Bounds-Based Rate-Adaptive Image Compressed Sensing Network

Yujun Huang, Bin Chen, Naiqi Li, Baoyi An, Shu‐Tao Xia, Yaowei Wang

TPAMI 2025 · Image generation › Image restoration · approach: Inverse imaging with learned priors · 9 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
add measurement-bounds-based rate-adaptive sampling instead of uniform sampling (ablation)MB-RACS · 7 datasets averagePSNR: average improvement of 1.89 dBImproved
set overall sparsity ratio from the sampling rate instead of fixed 0.15 (ablation)MB-RACS · 7 datasets averagePSNR: average improvement of 0.72 dBImproved
add decoder-side cross-iteration skip connections (ablation)MB-RACS · 7 datasets averagePSNR: average improvement of 0.09 dBImproved
increase sampling stages from 2 to 5 (ablation)MB-RACS multi-stage · 7 datasets averagePSNR: 0.30 dBImproved
increase sampling stages from 5 to 8 (ablation)MB-RACS multi-stage · 7 datasets averagePSNR: 0.05 dBImproved
increase sampling stages from 8 to 10 (ablation)MB-RACS multi-stage · 7 datasets averagePSNR: 0.004 dBNo clear effect
add distribution ratio optimization (ablation)MB-RACS multi-stage · 7 datasets averagePSNR: average improvement of 0.21 dBImproved
use MB-RACS_ss instead of single-stage saliency-based CASNetsingle-stage rate-adaptive CS · 7 datasets averagePSNR: 0.85dB/0.0182Improved
use measurement-bounds sampling instead of saliency-based sampling with ISTA-Net+ decoderrate-adaptive sampling · 6 datasetsPSNR vs uniform: saliency-based declines of 3.62 dB; ours std 0.21 dB vs 1.57 dBImproved
use MB-RACS_ss instead of OPINE-Net with uniform samplingcompressed sensing · 7 datasets averagePSNR/SSIM: 1.93dB/0.0175Improved
use MB-RACS_ss instead of CASNet_ssCASNet · CBSD68 (sampling rate 0.01)PSNR: 22.15 -> 23.81Improved

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