Abstract:
Attention to faces is linked to social and communicative developmental outcomes, but questions remain about when and why infants look to faces in their everyday lives. To...Show MoreMetadata
Abstract:
Attention to faces is linked to social and communicative developmental outcomes, but questions remain about when and why infants look to faces in their everyday lives. To make answering these questions tractable, advanced methodological techniques are needed. Thirty infants played with their parent while wearing head-mounted eye trackers at 4- and 8-months-old. Computer vision automatic face detection software (RetinaFace) identified faces, including facial features, from the first-person camera views of both the infant and the parent. Additionally, emotion recognition software identified moments when the parent smiled. Automatic face detection was highly convergent with human coding. We employed a Viterbi algorithm to ameliorate moments when a face was missed (i.e., false negative) or incorrectly detected (i.e., false positive). Finally, we compared the availability of and attention to faces across age. Faces were more available to 4-month-old dyads than 8-month-old dyads, and 4-month-old infants looked at faces more than 8-month-old infants. Infants at both ages primarily looked to the eye region of the face, and parent smiling did not impact attention to the face. Implications of the current work and future directions are discussed.
Date of Conference: 20-23 May 2024
Date Added to IEEE Xplore: 27 August 2024
ISBN Information: