⌁ Awesome World Model Hub Papers · Datasets · Projects
中文 GitHub
← Back to papers

PlayerOne: Egocentric World Simulator

arXiv 25.6 2025 63.7 method, system

TLDR

PlayerOne is the first egocentric realistic world simulator that generates egocentric videos aligned with user motion from an exocentric camera.

Reasoning

The paper introduces a novel egocentric world simulator with a coarse-to-fine training pipeline, part-disentangled motion injection, and joint 4D scene-video reconstruction, showing strong generalization. However, it lacks interactivity and reinforcement learning aspects, and the requirement of an exocentric camera limits applicability.

Read-first score

Read-first score 63.7, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 42.

Recency 8%
86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Methodology quality 25%
80

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=dataset,experiment,result

Topical relevance 42%
60

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 25%
46

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=dataset,github

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 178.

Keyword Scores

world simulator
10
video world model
9
generative world model
8
world model
7
world dynamics prediction
5
interactive world model
3
model-based reinforcement learning world model
0

Deep Analysis

Innovations

  • First egocentric realistic world simulator enabling immersive and unrestricted exploration in dynamic environments
  • Coarse-to-fine training pipeline: pretraining on large-scale egocentric text-video pairs for coarse understanding, then finetuning on synchronous motion-video data from egocentric-exocentric datasets via automatic construction
  • Part-disentangled motion injection scheme for precise control of part-level human movements
  • Joint reconstruction framework that progressively models both 4D scene and video frames to ensure scene consistency in long-form video generation

Methodology

PlayerOne uses a coarse-to-fine pipeline: first pretraining on large-scale egocentric text-video pairs for coarse-level egocentric understanding, then finetuning on synchronous motion-video data extracted from egocentric-exocentric video datasets via an automatic construction pipeline. It incorporates a part-disentangled motion injection scheme for precise part-level movement control and a joint reconstruction framework that progressively models both the 4D scene and video frames to maintain scene consistency in long-form generation.

Key Results

Experimental results demonstrate great generalization ability in precise control of varying human movements and world-consistent modeling of diverse scenarios, marking the first endeavor into egocentric real-world simulation.

Limitations

  • As the first work in egocentric real-world simulation, generalization to unseen scenarios or motion types may be limited
  • Reliance on synchronous egocentric-exocentric video data, which may be scarce or difficult to collect at scale

Resources

PaperProject

Tags