AI Paper Atlas

Research map / Detection and segmentation / CNN features

Joint Spectral Image Reconstruction and Semantic Segmentation with Cooperative Unfolding

Zijun He, Ping Wang, Xiaodong Wang, Chang Chen, Xin Yuan

CVPR 2026 · Detection and segmentation › CNN features · approach: Domain-specific segmentation and detection

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
feed the segmentation map into the reconstruction gradient descent step (ablation)CRSDUN-3stg baseline-1 · FVgNET simulationPSNR (dB): 37.33 -> 37.90Improved
add LISTA for gradient descent on the segmentation map (ablation)CRSDUN-3stg · FVgNET simulationPSNR (dB): 37.90 -> 38.03Improved
add LISTA for gradient descent on the segmentation map (ablation)CRSDUN-3stg · FVgNET simulationmIoU (%): 86.08 -> 86.55Improved
add cross-aggregation in the reconstruction network (not just segmentation) (ablation)CRSDUN-3stg · FVgNET simulationPSNR (dB): 38.60 -> 39.35Improved
add cross-aggregation in the segmentation network (not just reconstruction) (ablation)CRSDUN-3stg · FVgNET simulationmIoU (%): 85.49 -> 90.12Improved
increase unfolding stages from 3 to 5 (ablation)CRSDUN · FVgNET simulationmIoU (%): 90.11 -> 92.33Improved
increase unfolding stages from 5 to 7 (ablation)CRSDUN · FVgNET simulationmIoU (%): 92.33 -> 90.83Hurt
increase unfolding stages from 5 to 9 (ablation)CRSDUN · FVgNET simulationPSNR (dB): 39.88 -> 40.35Improved
use joint cooperative unfolding (CRSDUN-5stg) instead of two-stage SSR-9stg+Segreconstruction-then-segmentation · FVgNET simulationmIoU (%): 85.74 -> 92.33Improved
use CRSDUN-3stg instead of SSR-3stg+Segreconstruction-then-segmentation · FVgNET simulationPSNR (dB): 36.97 -> 39.35Improved

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