Effects of Clutter and Cueing Methods in AI-supported Visual Search

This study investigates how clutter and direct cueing methods interact to influence visual search performance in naturalistic aerial imagery. We conducted a mixed-design experiment comparing additive (bounding boxes) and subtractive cueing (contextual blur) across varying levels of visual clutter. Participants (N=20) completed three blocks of baseline search trials without support, followed by three blocks of AI-assisted trials. In each trial, they identified targets in aerial scenes and rated their perceived mental workload. Results indicate that both cueing methods performed equally well in simple environments, whereas subtractive cueing significantly reduced mental workload and reaction times when processing highly cluttered scenes. These findings emphasise the potential of subtractive cueing to optimise cognitive resource allocation in complex mSAR environments.