A View of Their Own: Capturing the Egocentric View of Infants and Toddlers with Head-Mounted Cameras

This method can help answer key questions in developmental psychology in the cognitive sciences such as what is the visual environment of infants and how does what they see influence their behavior? The main advantage of this technique is that it allows researchers to capture the view of an infant from their own perspective. This view is very different than that of an adult.

This method can also be applied to adults and other non-human model organisms. Researchers have attached cameras to the heads or backs of birds, bats, and other species in the wild. One challenge with this protocol is successfully placing the head camera.

Children can be fussy and sensitive to anything to anything touching their head so experimenter confidence is critical. Demonstrating this procedure will be graduate students Sarah Shroer and Catalina Suarez-Rivera. To begin ask a parent and the second experimenter to ensure that the child remains calm during the placement process and to distract the child as needed.

Then ask the parent to desensitize the infant to hand actions near her head by slightly touching the infant's hair several times. After this, lightly touch the infant's hair and head. Provide the parent with distracting toys and instruct the parent and the second experimenter to gently push the infants hand toward the toy.

While the infant is distracted place the head-mounted camera on the infant's head. Next, adjust the camera so that when the infant holds an object in front of his face the object is centered in the camera field of view. After this leave the room and begin the recording.

In the event that the camera is moved out of place or removed reenter the room to correct the camera. First outfit the infant with the head camera as previously described. Provide the parent and infant with several toys before leaving the room.

Subsample one frame of the video stream every five seconds. Then using software manually draw bounding boxes around each toy observed in the subsampled frame. If only part of an easily identifiable toy is visible in the frame draw a bounding box only around the visible part of the toy.

In this protocol head-mounted cameras were used to count the number of objects in the view of a parent-infant dyad. In these representative results the child had a greater number of frames with fewer objects in view compared to the parent. The proportion of the screen taken up by each object in view was also analyzed.

For both the parent and child there was a negative correlation between the number of objects in view and the visual sizes of those objects. Following this procedure further hand coding or computer vision algorithms can be used to answer more detailed questions about the properties of objects in the infant's view. This technique has been successfully used by our lab and others to address many research questions.

Our lab has had a great deal of success using head-mounted cameras to explore the natural home environment and visual experiences of infants across the globe.