Visual Search Strategies in Dense Virtual Reality Scenes Depend on Inter-Object Spacing
Abstract
Visual search in virtual reality (VR) relies on the coordination of peripheral vision, eye movements, and head movements. In dense immersive scenes, inter-object spacing is expected to influence search behavior by modulating local visual interference, while also changing the angular extent of the explored scene. In this paper, we investigate how inter-object spacing shapes visual search strategies in VR using ecologically valid textured 3D objects. We conducted a user study in which participants searched for target objects among distractors under three spacing conditions, while keeping the number of objects constant and recording both eye and head movements. Conventional visual-search measures, such as search time, fixation count, and mean fixation duration, are typically expected to vary consistently with task difficulty. However, our results reveal a mixed pattern: search time and fixation count increase with spacing, whereas mean fixation duration decreases. This apparent paradox highlights the limitations of standard metrics in VR, where performance is strongly influenced by scene extent and head contribution. To address this, we complement traditional measures with detailed scanpath analyses. We show that dense layouts induce more refixations, more frequent direction changes, and less stable exploration patterns. Although absolute scanpath length increases with spacing, normalization by inter-object distance reveals that the relative exploration cost is highest in the densest condition. In contrast, wider spacing is associated with a stronger contribution of head movements to the search process. Taken together, these findings show that inter-object spacing does not simply make search globally easier or harder, but reorganizes visual exploration strategies in immersive VR in a way that is consistent with visual crowding principles, while also reflecting other spacing-dependent factors such as scene extent, effective field of view, and head-movement demands.