Papers › Joint Learning of Blind Super-Resolution and Crack Segmentation for Realistic Degraded Images

Joint Learning of Blind Super-Resolution and Crack Segmentation for Realistic Degraded Images

24 Feb 2023IEEE Transactions on Instrumentation and Measurement (TIM) 2024 3arXiv:2302.12491archive 2025-07-28

Yuki Kondo, Norimichi Ukita

This paper proposes crack segmentation augmented by super resolution (SR) with deep neural networks. In the proposed method, a SR network is jointly trained with a binary segmentation network in an end-to-end manner. This joint learning allows the SR network to be optimized for improving segmentation results. For realistic scenarios, the SR network is extended from non-blind to blind for processing a low-resolution image degraded by unknown blurs. The joint network is improved by our proposed two extra paths that further encourage the mutual optimization between SR and segmentation. Comparative experiments with State of The Art (SoTA) segmentation methods demonstrate the superiority of our joint learning, and various ablation studies prove the effects of our contributions.

PaperPDFConference PDFCode

Code

yuki-11/csbsr officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Blind Super-ResolutionCrack SegmentationSegmentationSuper-Resolution

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet) AHD95 22.52 #1 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet) Average IOU 0.552 #1 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet) HD95_min 20.92 #1 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet) IoU_max 0.573 #1 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW) AHD95 21.70 #2 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW) Average IOU 0.551 #2 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW) HD95_min 18.73 #2 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW) IoU_max 0.573 #2 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ HRNet+OCR) AHD95 20.29 #4 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ HRNet+OCR) Average IOU 0.534 #4 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ HRNet+OCR) HD95_min 17.54 #4 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ HRNet+OCR) IoU_max 0.553 #4 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW+BlurSkip) AHD95 19.10 #5 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW+BlurSkip) Average IOU 0.528 #5 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW+BlurSkip) HD95_min 18.06 #5 of 7 Archive leaderboard report
Crack Segmentation khanhha's dataset - 4x upscaling (blind) CSBSR (w/ PSPNet+FOW+BlurSkip) IoU_max 0.550 #5 of 7 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections