[ Abstract ]

In modern indoor space, not everyone can enjoy windows with good views. Moreover, extended indoor hours and windowless spaces can hamper occupant physical psychological health. Conventional solutions, like digital windows, face issues with fixed viewpoints and high cost of installation. Meanwhile, advanced XR devices have made real- time tracking and high-fidelity rendering feasible, enabling the creation of virtual windows that replicate and enhance the experience of physical windows. This paper proposes a framework for designing XR-based vir- tual windows aimed at enhancing indoor spatial experience. The frame- work consists of three core components: (1) Window Position Analyzer, which calculates feasible window placement by analyzing the physical environment and user movements; (2) Window Configuration Manager, which defines parameters for the window frame and glazing; and (3) Vir- tual Scenery Generator, which creates immersive views through object arrangement and motion simulation. By synthesizing recent advance- ments in computer graphics, sensing technologies, and neural rendering, this study discusses how to achieve naturalistic scenes and realistic light- ing effects without major architectural modifications. We then present a proof-of-concept scenario focusing on residential contexts, highlighting the key technical considerations and design choices required. The findings illustrate that XR virtual windows can provide a flexible, personalized, and psychologically beneficial alternative to traditional windows, repre- senting a promising direction for the future of indoor space experience.

Keywords: Extended Reality (XR) · Virtual Window · Indoor Environment · Scene Understanding · Neural Rendering · Generative AI