UCM: Unifying Camera Control and Memory with Time-aware Positional Encoding Warping for World Models

TL;DR

UCM introduces a unified framework combining long-term memory and camera control in world models, leveraging time-aware positional encoding and efficient diffusion transformers to improve scene consistency and controllability in video generation.

Contribution

The paper proposes UCM, a novel approach that unifies memory and camera control using time-aware encoding and scalable data curation, advancing scene consistency and controllability.

Findings

Outperforms state-of-the-art in long-term scene consistency

Achieves precise camera control in high-fidelity video generation

Efficient dual-stream diffusion transformer reduces computational overhead

Abstract

World models based on video generation demonstrate remarkable potential for simulating interactive environments but face persistent difficulties in two key areas: maintaining long-term content consistency when scenes are revisited and enabling precise camera control from user-provided inputs. Existing methods based on explicit 3D reconstruction often compromise flexibility in unbounded scenarios and fine-grained structures. Alternative methods rely directly on previously generated frames without establishing explicit spatial correspondence, thereby constraining controllability and consistency. To address these limitations, we present UCM, a novel framework that unifies long-term memory and precise camera control via a time-aware positional encoding warping mechanism. To reduce computational overhead, we design an efficient dual-stream diffusion transformer for high-fidelity generation. Moreover, we introduce a scalable data curation strategy utilizing point-cloud-based rendering to simulate scene revisiting, facilitating training on over 500K monocular videos. Extensive experiments on real-world and synthetic benchmarks demonstrate that UCM significantly outperforms state-of-the-art methods in long-term scene consistency, while also achieving precise camera controllability in high-fidelity video generation.