Papers

MAVE:从电影镜头自动重建可编辑 3D 布景草稿

MAVE: Reconstructing Editable 3D Blockouts from Film Shots

Why we picked it

用 Agent 从电影片段直接生成可编辑的 Blender 布景,一个镜头成本才 0.054 美元,比手工重建快一倍还多。做预览、分镜或给视频生成模型做条件控制的人都该看看,代码和 3K 数据集都要开源。

研究团队提出 MAVE 框架,能从普通电影片段自动生成可编辑的 3D 布景草稿,并配套发布了评测基准 MAVEBench,包含 50 个电影镜头和 416 个关键帧,用于评估相机运动、空间布局和人物走位。传统做法靠人工从参考视频重建布景,成本高,这套框架把流程自动化了。

Body

技术上,MAVE 分三步:先在统一的度量坐标系里恢复相机参数、人物动作和几何证据;再由 Agent 在这些观测的约束下搭建可编辑场景;最后通过程序化检查报告不一致之处并迭代修正。最终输出可直接编辑的 Blender 场景和与原视频逐帧对齐的布览视频。在 MAVEBench 上,三个 Agent 的整体分数比各自的 vanilla 基线提升 27.5 到 41.1 分;搭配 DeepSeek-1017 达到 73.2 分,单个镜头平均 API 成本仅 0.054 美元,重建总耗时减少 53%。

团队同时构建了 MAVEData-3K 数据集,含 3,000 组布景-视频配对样本,将公开发布,可用于训练视频生成模型。创作者可在 Blender 中改变视角和人物路径,渲染编辑后的场景来控制视频生成条件。代码已在 GitHub 上开源。

Original · arXiv: DeepSeek

MAVE: Reconstructing Editable 3D Blockouts from Film Shots

Animated 3D blockouts support previsualization and provide structural conditions for video generation, but manually reconstructing them from reference videos is costly. This limits both their use by creators and the production of paired training data. We introduce MAVEBench, a benchmark of 50 film shots and 416 keyframes for evaluating camera motion, spatial layout, and character blocking. We propose MAVE, a video-to-blockout framework combining Metric observation, Agentic modeling, and evidence-based VErification. It first recovers cameras, human motion, and geometric evidence in a shared metric coordinate system. An agent then constructs an editable environment under constraints from these observations, while programmatic checks report inconsistencies to guide refinement. The outputs are editable Blender scenes and frame-aligned blockout videos. On MAVEBench, MAVE improves overall scores by 27.5-41.1 points across three agents over their corresponding vanilla baselines. With DeepSeek-1017, it achieves a score of 73.2 at an average API cost of only USD 0.054 per shot, outperforming the strongest vanilla baseline while reducing total reconstruction time by 53%. We also construct MAVEData-3K, a dataset of 3,000 blockout-video pairs for model training, and will release it publicly. Creators can change viewpoints and character paths, then render the edited scenes to condition video generation. Code is available at https://github.com/yfChang-cv/MAVE.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include MAVE, Blender, DeepSeek, video generation.

Primary source verification: published directly by arXiv: DeepSeek with first-party authority.