Problem
Research questions and friction points this paper is trying to address.
video restoration
temporal inconsistency
output reliability
physical trace
degradation
Revisiting the Current Frame: Physical-Trace-Guided Network Output Correction for Video Restoration
🤖 AI Summary
This work addresses the challenge of inconsistent spatial reliability in video restoration, which arises from physical imaging discrepancies, occlusions, or imperfect temporal alignment. To tackle this issue, the authors propose ANCHOR, a novel framework that treats the low-quality current frame as a temporal alignment anchor and leverages physical trace evidence to estimate a spatial confidence field. This confidence field adaptively fuses the network-restored output with the original observation. ANCHOR introduces, for the first time, a physical-trace-guided spatial confidence mechanism, establishing a model-agnostic, reliability-aware correction framework that significantly enhances spatial consistency in restoration results. Experiments on HDR video reconstruction and video deraining demonstrate that ANCHOR consistently boosts the performance of diverse state-of-the-art restoration models, validating its generality and effectiveness.
📝 Abstract
Video restoration methods exploit temporal information to recover information missing from degraded observations. However, reference frames within the sequence may introduce inconsistent degradation, content discrepancy, or reconstruction errors due to physical image-formation variations, occlusion, and imperfect temporal aggregation. Existing approaches mainly focus on improving restoration networks, while the reliability of the generated outputs at different spatial locations remains largely unexplored. In this work, we propose ANCHOR, a model-agnostic framework that revisits the low-quality current frame as a temporally aligned anchor for video restoration correction. Specifically, ANCHOR estimates a spatial trust field from heterogeneous physical-trace evidence and adaptively balances the restoration proposal with the original observation. Experiments on High Dynamic Range video reconstruction and video deraining demonstrate consistent improvements across various state-of-the-art restoration models, validating the effectiveness of reliability-aware output correction for video restoration.
Innovation
Methods, ideas, or system contributions that make the work stand out.
physical-trace-guided
output correction
spatial trust field
model-agnostic
video restoration
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
