Abstract
Purpose: :
Depth perception is an integral part of being able to correctly interpret one’s environment, particularly for mobility or grasping tasks. The issue of whether depth cues can be appreciated with monocular camera-based artificial vision devices is an unanswered question. The purpose of this pilot study was to determine whether subjects using the BrainPort vision device would be able to interpret spatial relationships between two objects.
Methods: :
4 blindfolded sighted adults (1 male; 3 females) participated in this study. All had previous exposure to the BrainPort device. Experiments were conducted with both standing and seated tasks with a fixed camera field of view (73 and 40 degrees respectively). The standing tasks used a 13 x 11 inch black box on a white tile floor. The tasks were to 1) determine whether the box was 5, 10 or 20 feet away 2) determine the distance away (five foot increments) until the object could no longer be detected 3) determine the separation distance (6 inch increments) between two boxes 4) estimate the distance of the box placed at random distances. Tabletop conditions were completed on a black felt backdrop using 1.5 inch cylinders, a 12x4 inch felt rectangle and a felt circle with a diameter of 5.5 inches. The tasks included: A) moving two cylinders horizontally closer together (from 6 inches) until separation could no longer be appreciated B) determining whether a felt circle was above or below a felt rectangle C) estimate the distance (in inches) between the two cylinders D) determine whether the felt circle was in front or behind the rectangle.
Results: :
For standing tasks subjects 1) averaged 66% correct in estimating the distance to the box, 2) could detect the presence of the box from an average 18.75 feet, 3) determined the two box discrimination distance to be an average of 9 inches 4) could estimate boxes placed at random distances within 1.75 feet error. For the seated tasks, subjects A) could detect the separation between two cylinders with average 2.75 inches minimal distance, B) could determine the spatial relationship between felt shapes with 79% accuracy, C) could estimate the distance between cylinders with at least 0.5 inches distance, D) could determine the spatial relationship between overlapping shapes with 83% accuracy.
Conclusions: :
This study demonstrates that it is possible to discern crude depth information using a two dimensional tactile stimulus on the tongue. It is likely that training for this specific task, as well as increased resolution would enhance this ability. Methods to improve depth appreciation are needed in order to enable mobility and grasping tasks with artificial vision devices.
Keywords: low vision • depth • perception