Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

$\text{F}^2\text{HDR}$: Two-Stage HDR Video Reconstruction via Flow Adapter and Physical Motion Modeling

Huanjing Yue, Dawei Li, Shaoxiong Tu, Jingyu Yang

Abstract

Reconstructing High Dynamic Range (HDR) videos from sequences of alternating-exposure Low Dynamic Range (LDR) frames remains highly challenging, especially under dynamic scenes where cross-exposure inconsistencies and complex motion make inter-frame alignment difficult, leading to ghosting and detail loss. Existing methods often suffer from inaccurate alignment, suboptimal feature aggregation, and degraded reconstruction quality in motion-dominated regions. To address these challenges, we propose $\text{F}^2\text{HDR}$, a two-stage HDR video reconstruction framework that robustly perceives inter-frame motion and restores fine details in complex dynamic scenarios. The proposed framework integrates a flow adapter that adapts generic optical flow for robust cross-exposure alignment, a physical motion modeling to identify salient motion regions, and a motion-aware refinement network that aggregates complementary information while removing ghosting and noise. Extensive experiments demonstrate that $\text{F}^2\text{HDR}$ achieves state-of-the-art performance on real-world HDR video benchmarks, producing ghost-free and high-fidelity results under large motion and exposure variations.

$\text{F}^2\text{HDR}$: Two-Stage HDR Video Reconstruction via Flow Adapter and Physical Motion Modeling

Abstract

Reconstructing High Dynamic Range (HDR) videos from sequences of alternating-exposure Low Dynamic Range (LDR) frames remains highly challenging, especially under dynamic scenes where cross-exposure inconsistencies and complex motion make inter-frame alignment difficult, leading to ghosting and detail loss. Existing methods often suffer from inaccurate alignment, suboptimal feature aggregation, and degraded reconstruction quality in motion-dominated regions. To address these challenges, we propose , a two-stage HDR video reconstruction framework that robustly perceives inter-frame motion and restores fine details in complex dynamic scenarios. The proposed framework integrates a flow adapter that adapts generic optical flow for robust cross-exposure alignment, a physical motion modeling to identify salient motion regions, and a motion-aware refinement network that aggregates complementary information while removing ghosting and noise. Extensive experiments demonstrate that achieves state-of-the-art performance on real-world HDR video benchmarks, producing ghost-free and high-fidelity results under large motion and exposure variations.
Paper Structure (15 sections, 16 equations, 6 figures, 5 tables)

This paper contains 15 sections, 16 equations, 6 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Stage I: Coarse Fusion Stage
  5. Flow Estimation with Flow Adapter
  6. Physical Motion Modeling
  7. Stage II: Refinement Stage
  8. Motion Guided Feature Modulation
  9. Feature Enhancement and Fusion
  10. Training Loss
  11. Experiments
  12. Experimental Setup
  13. Comparisons with State-of-the-art
  14. Ablation Study
  15. Conclusion

Figures (6)

  • Figure 1: Comparison on the real-world HDR video reconstruction benchmark shu2024towards with state-of-the-art methods xu2024hdrflowcui2024exposure. The challenging scene involves both global and local motion, while the reference frames suffer from severe information loss. Our $\text{F}^2\text{HDR}$ achieves the most robust optical flow estimation and the most accurate detail restoration.
  • Figure 2: Network architecture of the proposed $\text{F}^2\text{HDR}$, which includes coarse fusion stage and refinement stage.
  • Figure 3: Qualitative comparisons with state-of-the-art methods on DeepHDRVideo dataset chen2021hdr.
  • Figure 4: Qualitative comparisons with state-of-the-art methods on Real-HDRV dataset shu2024towards.
  • Figure 5: Qualitative Comparisons of optical flow and warped LDR on DeepHDRVideo (Rows 1-2) chen2021hdr and Real-HDRV (Rows 3-4) shu2024towards Datasets.
  • ...and 1 more figures