New benchmark and PECA module tackle heterogeneous stereo deblurring

Researchers have introduced a new benchmark and method for addressing heterogeneous stereo deblurring, a problem common in smartphone cameras where different modules can cause asymmetric blur. The proposed solution includes the Heterogeneous Stereo Deblurring (HSD) dataset, built from real smartphone captures, and a novel module called Physically- and Epipolar-constrained Cross Attention (PECA). PECA uses physically derived constraints to guide feature fusion across camera views, improving deblurring performance and efficiency for various existing models like CNNs, Transformers, and NAFNet. AI

IMPACT Introduces a novel approach to improve image quality in stereo vision systems, potentially impacting mobile photography and AR applications.

RANK_REASON Academic paper introducing a new dataset and method for a specific computer vision task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• CNN
• NAFNet
• PECA
• transformer

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →