Lune

CVPR Findings2026

Text-Driven Reasoning Video Editing via Reinforcement Learning on Digital Twin Representations

Yiqing Shen, Chenjia Li, Mathias Unberath

Abstract

Text-driven video editing enables users to modify video content only using text queries. While existing methods can modify video content if explicit descriptions of editing targets with precise spatial locations and temporal boundaries are provided, these requirements become impractical when users attempt to conceptualize edits through implicit queries referencing semantic properties or object relationships. We introduce reasoning video editing, a task where video editing models must interpret implicit queries through multi-hop reasoning to infer editing targets before executing modifications, and a first model attempting to solve this complex task, RIVER (Reasoning-based Implicit Video Ed-itoR). RIVER decouples reasoning from generation through digital twin representations of video content that preserve spatial relationships, temporal trajectories, and semantic attributes. A large language model then processes this representation jointly with the implicit query, performing multi-hop reasoning to determine modifications, then outputs structured instructions that guide a diffusion-based editor to execute pixel-level changes. RIVER training uses reinforcement learning with rewards that evaluate reasoning accuracy and generation quality. Finally, we introduce RVEBenchmark, a benchmark of 100 videos with 519 implicit queries spanning three levels and categories of reasoning complexity specifically for reasoning video editing. RIVER demonstrates best performance on the proposed RVEBenchmark and also achieves state-of-the-art performance on two additional video editing benchmarks (VegGIE and FiVE), where it surpasses six baseline methods.

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

  • Depth Anything V2 cites Segment Anything
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on19

  1. LoRA: Low-Rank Adaptation of Large Language ModelsEdward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu et al.ICLR 2022 · 18,833 citations
  2. Segment AnythingAlexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao et al.ICCV 2023 · 13,211 citations
  3. Depth Anything V2Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao et al.NeurIPS 2024 · 2,305 citations
  4. MUSIQ: Multi-scale Image Quality TransformerJunjie Ke, Qifei Wang, Yilin Wang, Peyman Milanfar et al.ICCV 2021 · 1,325 citations
  5. CLIPScore: A Reference-free Evaluation Metric for Image CaptioningJack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras et al.EMNLP 2021 · 937 citations

Related papers

  1. VEGGIE: Instructional Editing and Reasoning Video Concepts with Grounded GenerationShoubin Yu, Difan Liu, Ziqiao Ma, Yicong Hong et al.ICCV 2025 · 2 citations
  2. CoT-Edit: Let CoT Guide Instruction Video EditingSen Liang, Fengbin Guan, Youliang Zhang, Xin Li et al.CVPR 2026 · 5 citations
  3. Online Reasoning Video Segmentation with Just-in-Time Digital TwinsYiqing Shen, Bohan Liu, Chenjia Li, Lalithkumar Seenivasan et al.ICCV 2025 · 7 citations
  4. Editing the Moving World: Model Editing for Video LLMsQian Zhang, Xinye Li, Xiaokai Wu, Junhao Xu et al.ACL 2026
  5. Are Image-to-Video Models Good Zero-Shot Image Editors?Zechuan Zhang, Zhenyuan Chen, Zongxin Yang, Yi YangCVPR 2026 · 4 citations
Dusk over the sea between two cliffs drawn in fine vertical lines