Saved in:
Bibliographic Details
Main Authors: Kara, Ozgur, Chen, Yujia, Yang, Ming-Hsuan, Rehg, James M., Chu, Wen-Sheng, Tran, Du
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.20809
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • We present Split-then-Merge (StM), a novel framework designed to enhance control in generative video composition and address its data scarcity problem. Unlike conventional methods relying on annotated datasets or handcrafted rules, StM splits a large corpus of unlabeled videos into dynamic foreground and background layers, then self-composes them to learn how dynamic subjects interact with diverse scenes. This process enables the model to learn the complex compositional dynamics required for realistic video generation. StM introduces a novel transformation-aware training pipeline that utilizes a multi-layer fusion and augmentation to achieve affordance-aware composition, alongside an identity-preservation loss that maintains foreground fidelity during blending. Experiments show StM outperforms SoTA methods in both quantitative benchmarks and in humans/VLLM-based qualitative evaluations. More details are available at our project page: https://split-then-merge.github.io