Purchase this article with an account.
Sandra Lafortune, Caroline Blais, Justin Duncan, Amanda Estephan, Daniel Fiset; The importance of the natural contour for visual feature integration in face processing.. Journal of Vision 2015;15(12):159. doi: https://doi.org/10.1167/15.12.159.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Hills and coll. (2014) have suggested that the natural face contour is a fundamental part of the whole face. Here, we investigated the impact of the natural contour on face recognition by measuring the integration index (Φ), a measure of feature integration in face processing (Gold et al., 2012). Twelve participants (8 women) completed 840 trials per day (divided into three blocks) for 3 consecutive days. The experiment was composed of seven conditions based on the features available for the task: I) right eye, II) left eye, III) nose, IV) mouth, V) all former features (as in Gold et al.), VI) facial contour and VII) whole face. On each trial, a target stimulus from one of the seven conditions (randomly intermixed across trials) was briefly presented (500 ms), and followed by the simultaneous presentation of six stimuli from the same condition. The task was to identify which of the six was the same as the one previously presented. For each condition, contrast of the target stimulus was adjusted using QUEST (Watson & Pelli, 1983) to reach an accuracy threshold of 75%. Performance, as indexed by the contrast at which each participant reached the threshold level (provided by QUEST at the end of each block of day 2 and 3), was significantly better with contour than without (t(11)= 4.14, p= 0.002). This observation can be explained by the fact that the former contains more information than the latter. Most importantly, the integration index was significantly higher for the condition with (Φ= 1.39) than without (Φ= 0.64) contour (t(11)= -3.2, p= 0.008). These results suggest that presence of the natural face contour is important for efficient integration during face processing.
Meeting abstract presented at VSS 2015
This PDF is available to Subscribers Only