Lune

CVPR2026Top-tier venue

F^2HDR: 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 F 2 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 F 2 HDR achieves state-of-the-art performance on realworld HDR video benchmarks, producing ghost-free and high-fidelity results under large motion and exposure variations. Our code is available at https://github.com/ wei1895/F2HDR.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • LAN-HDR: Luminance-based Alignment Network for High Dynamic Range Video Reconstruction cites HDR Video Reconstruction: A Coarse-to-fine Network and A Real-world Benchmark Dataset
  • HDRFlow: Real-Time HDR Video Reconstruction with Large Motions cites HDR Video Reconstruction: A Coarse-to-fine Network and A Real-world Benchmark Dataset
  • HDRFlow: Real-Time HDR Video Reconstruction with Large Motions cites LAN-HDR: Luminance-based Alignment Network for High Dynamic Range Video Reconstruction
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on12

  1. HDR Video Reconstruction: A Coarse-to-fine Network and A Real-world Benchmark DatasetGuanying Chen, Chaofeng Chen, Shi Guo, Zhetong Liang et al.ICCV 2021 · 70 citations
  2. Alignment-free HDR Deghosting with Semantics Consistent TransformerSteven Tel, Zongwei Wu, Yulun Zhang, Barthélémy Heyrman et al.ICCV 2023 · 44 citations
  3. LAN-HDR: Luminance-based Alignment Network for High Dynamic Range Video ReconstructionHaesoo Chung, Nam Ik ChoICCV 2023 · 20 citations
  4. Progressive and Selective Fusion Network for High Dynamic Range ImagingQian Ye, Jun Xiao, Kin-Man Lam, Takayuki OkataniACM MM 2021 · 19 citations
  5. HDRFlow: Real-Time HDR Video Reconstruction with Large MotionsGangwei Xu, Yujin Wang, Jinwei Gu, Tianfan Xue et al.CVPR 2024 · 12 citations

Related papers

  1. DeAltHDR: Learning HDR Video Reconstruction from Degraded Alternating Exposure SequencesShuohao Zhang, Zhilu Zhang, Rongjian Xu, Xiaohe Wu et al.ICLR 2026
  2. Exposure Completing for Temporally Consistent Neural High Dynamic Range Video RenderingJiahao Cui, Wei Jiang, Zhan Peng, Zhiyu Pan et al.ACM MM 2024 · 3 citations
  3. Towards Real-World HDR Video Reconstruction: A Large-Scale Benchmark Dataset and A Two-Stage Alignment NetworkYong Shu, Liquan Shen, Xiangyu Hu, Mengyao Li et al.CVPR 2024 · 8 citations
  4. LRHDR: Learning Representation-enhanced HDR Video ReconstructionChenzhuo Liao, Xin Chen, Bingchen Li, Yu Meng et al.CVPR 2026
  5. Mono4DGS-HDR: High Dynamic Range 4D Gaussian Splatting from Alternating-exposure Monocular VideosJinfeng Liu, Lingtong Kong, Mi Zhou, Jinwei Chen et al.ICLR 2026 · 3 citations
Dusk over the sea between two cliffs drawn in fine vertical lines