![[2602.18981] How Far Can We Go with Pixels Alone? A Pilot Study on Screen-Only Navigation in Commercial 3D ARPGs](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.18981] How Far Can We Go with Pixels Alone? A Pilot Study on Screen-Only Navigation in Commercial 3D ARPGs
Summary
This study explores the effectiveness of screen-only navigation in 3D ARPGs, demonstrating how visual affordances can guide gameplay, while also highlighting limitations in current models.
Why It Matters
Understanding navigation in complex 3D environments is crucial for game design and AI development. This research provides a baseline for evaluating visual navigation systems, emphasizing the need for further exploration in AI-driven game navigation.
Key Takeaways
- The study introduces a navigation agent that relies solely on visual inputs.
- Pilot experiments indicate the agent can navigate effectively but has limitations.
- The research highlights the need for improved models for comprehensive navigation.
- Visual affordances play a significant role in guiding player movement in games.
- The findings call for more attention to visual navigation tasks in AI research.
Computer Science > Artificial Intelligence arXiv:2602.18981 (cs) [Submitted on 21 Feb 2026] Title:How Far Can We Go with Pixels Alone? A Pilot Study on Screen-Only Navigation in Commercial 3D ARPGs Authors:Kaijie Xu, Mustafa Bugti, Clark Verbrugge View a PDF of the paper titled How Far Can We Go with Pixels Alone? A Pilot Study on Screen-Only Navigation in Commercial 3D ARPGs, by Kaijie Xu and 2 other authors View PDF HTML (experimental) Abstract:Modern 3D game levels rely heavily on visual guidance, yet the navigability of level layouts remains difficult to quantify. Prior work either simulates play in simplified environments or analyzes static screenshots for visual affordances, but neither setting faithfully captures how players explore complex, real-world game levels. In this paper, we build on an existing open-source visual affordance detector and instantiate a screen-only exploration and navigation agent that operates purely from visual affordances. Our agent consumes live game frames, identifies salient interest points, and drives a simple finite-state controller over a minimal action space to explore Dark Souls-style linear levels and attempt to reach expected goal regions. Pilot experiments show that the agent can traverse most required segments and exhibits meaningful visual navigation behavior, but also highlight that limitations of the underlying visual model prevent truly comprehensive and reliable auto-navigation. We argue that this system provides a concrete...