Cross-Stage Patch Interaction for Coarse-to-Fine Image Deblurring
- 1School of Computing, Gachon University, Seongnam-si 13120, Republic of Korea
Abstract
Real-world motion blur exhibits complex, spatially varying characteristics that challenge existing deblurring methods. While multi-stage architectures offer progressive refinement, simple inter-stage interactions often limit effective information transfer. To address this, we propose CPINet, a Cross-stage Patch Interaction Network for coarse-guided image deblurring. CPINet employs a two-stage coarse-to-fine framework: the first stage uses an Uncertain Flow-based Prior Network (UFPNet) for coarse restoration, followed by a second stage for fine-grained patch-level refinement. The second stage utilizes a Coarse-guided Patch Encoding (CPE) block with a Patch-wise Cross-Channel Attention Gating (PCAG) mechanism to integrate first-stage features. Furthermore, a Patch-based Multi-Kernel Dual-Branch (P-MKDB) block is incorporated into the decoder to effectively handle diverse local blur patterns. Experiments on GoPro, HIDE, and RealBlur show competitive deblurring performance, while ablation studies support the contribution of the proposed cross-stage patch interaction and patch-wise refinement. Notably, CPINet improves the mean PSNR from 36.98 dB to 37.61 dB and the mean SSIM from 0.954 to 0.961 relative to UFPNet, with both metrics averaged across RealBlur-J and RealBlur-R. These improvements demonstrate its effectiveness in challenging real-world scenarios.
Related
- A novel inertial viscosity Z-algorithm for common fixed point problems with an application to image deblurring
- Marginalization-Based Regularization for Separable Nonlinear Least Squares Problems with Applications to Semi-Blind Deblurring
- Dynamic analysis and spatiotemporal evolution of multi-stage knowledge dissemination under AI-enabled dual-channel “Interpersonal–Human-Machine” interaction
- A One-Stage Approach for Estimating a Meta-Analytic Continuous-Time Random-Intercept Cross-Lagged Panel Model
- Self-compassion, psychological distress, and mental well-being in a predominantly young student sample: Cross-sectional associations and continuous interaction analysis
BibTeX
@article{yoo2026crossstage,
title = {{Cross-Stage Patch Interaction for Coarse-to-Fine Image Deblurring}},
author = {Sunghwan Yoo and Seokkyeong Boo and Yong Ju Jung},
journal = {Mathematics},
year = {2026},
month = oct,
doi = {10.3390/math14203808},
url = {https://www.mdpi.com/2227-7390/14/20/3808},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Cross-Stage Patch Interaction for Coarse-to-Fine Image Deblurring
AU - Sunghwan Yoo
AU - Seokkyeong Boo
AU - Yong Ju Jung
PY - 2026
DA - 2026/10/20
JO - Mathematics
DO - 10.3390/math14203808
UR - https://www.mdpi.com/2227-7390/14/20/3808
AB - Real-world motion blur exhibits complex, spatially varying characteristics that challenge existing deblurring methods. While multi-stage architectures offer progressive refinement, simple inter-stage interactions often limit effective information transfer. To address this, we propose CPINet, a Cross-stage Patch Interaction Network for coarse-guided image deblurring. CPINet employs a two-stage coarse-to-fine framework: the first stage uses an Uncertain Flow-based Prior Network (UFPNet) for coarse restoration, followed by a second stage for fine-grained patch-level refinement. The second stage utilizes a Coarse-guided Patch Encoding (CPE) block with a Patch-wise Cross-Channel Attention Gating (PCAG) mechanism to integrate first-stage features. Furthermore, a Patch-based Multi-Kernel Dual-Branch (P-MKDB) block is incorporated into the decoder to effectively handle diverse local blur patterns. Experiments on GoPro, HIDE, and RealBlur show competitive deblurring performance, while ablation studies support the contribution of the proposed cross-stage patch interaction and patch-wise refinement. Notably, CPINet improves the mean PSNR from 36.98 dB to 37.61 dB and the mean SSIM from 0.954 to 0.961 relative to UFPNet, with both metrics averaged across RealBlur-J and RealBlur-R. These improvements demonstrate its effectiveness in challenging real-world scenarios.
SN - 2227-7390
ER -