Abstract
Humans are able to search for a target in cluttered scenes more easily when it contrasts with its background. However, the way in which humans combine different visual cues to detect a pop-out target is not fully understood, particularly for peripheral viewing in displays with partial occlusions and clutter. Here, we focus on surface reflectance and motion. Inspired by so-called “dead-leaves stimuli”, we present an array of ellipses with different surface reflectance (luminance and color) and motion features, and define a target as a group of spatially-localized leaves with similar visual properties as the background leaves except for the cues of interest. A trial could contain a target defined by a single cue, a target defined by two cues jointly, or two targets each defined by a single cue, leading to single-cue, cue-combination, or cue-conflict conditions. Participants clicked on the center of the target with a mouse if they detected it, or right clicked if they detected nothing. A pre-experiment using a staircase procedure was used to titrate the cue difficulty before the formal experiment. If humans independently combine the two cues in the detection task, the cue-conflict condition should be as easy as the cue-combination condition, and better than single-cue conditions. Alternatively, if humans sum information from both cues for detection, the cue-combination but not the cue-conflict condition should outperform single-cue conditions. Results showed that the presence of both cues lead to improved performance compared with single-cue conditions, with better detectability in the cue-combination than the cue-conflict condition. However, given that participants could detect either of the two targets in the cue-conflict condition, the presence of both cues also drove additional accuracy improvement than expected by pure summation. The combination of surface reflectance and motion is overall better explained by a summation model than an independent-combination model.