EyeDraw: A System for Drawing Pictures with Eye Movements
Anthony Hornof, Anna Cavender, Rob Hoselton · 2004 · Proceedings of the 6th International ACM SIGACCESS Conference on Computers and Accessibility (Assets 04) · doi:10.1145/1028630.1028647
Summary
This paper presents EyeDraw, a software program that enables children with severe mobility impairments to draw pictures using only their eye movements tracked by an LC Technologies Eyegaze eye tracker. The system is designed to provide creative developmental experiences equivalent to those nondisabled children have with crayons and drawing programs. The key design innovation is a two-state interaction model that separates "looking" from "drawing": a green eye cursor indicates the Looking state where the user can freely view their picture without making marks, and a red cursor indicates the Drawing state where eye movements create lines. The transition between states is controlled by dwell time — holding the gaze steady for 500 milliseconds (adjustable) triggers the Drawing state, and moving the eyes away before a second 500ms dwell cancels the action. This approach solves the Midas touch problem that plagued earlier free-eye drawing systems where ink was deposited everywhere the user looked, making it impossible to examine the drawing without altering it. EyeDraw uses a CPM-GOMS (Cognitive, Perceptual, and Motor) analysis to understand the interaction between device, perceptual, cognitive, and oculomotor processes, which informed design decisions about dwell times and auditory feedback. The system includes line and circle drawing tools, a dot grid for spatial reference, undo functionality, and save/retrieve capabilities, all controllable entirely by eye gaze.
Key findings
User testing with 10 nondisabled participants (4 female, 6 male; ages 7-16 with average age 12) demonstrated that children with no prior eye tracking experience could learn to use EyeDraw within a single session lasting about an hour. The ideal dwell time was found to be approximately 500ms, representing a careful balance — shorter times risk unintended commands while longer times slow interaction. Participants generally rated the program easy to learn (mean 1.6 on a 1-5 scale, lower being easier) and easy to use (1.7), though controlling the eye cursor (2.2) and saving drawings (1.3) varied in difficulty. Four of the participants produced drawings with lines and circles that could be identified as recognizable pictures (a girl, a house, a car, a butterfly), suggesting the drawings corresponded to the "controlled scribble" and "basic forms" stages of children's drawing development. The youngest participant (age 7) did not produce a full picture but gained control over the drawing process during the session. The CPM-GOMS analysis revealed that auditory feedback (an audible click confirming drawing commands) is critical because it is perceived roughly 30ms faster than visual feedback, allowing users to begin planning their next movement sooner. A user observation study with disabled users of the Eyegaze Communication System was underway at the time of writing, with one power user already providing suggestions including color tools and integration into the main Eyegaze menu.
Relevance
This research is significant for accessibility practice because it expands the scope of eye tracking beyond communication and basic computer control into creative expression, an area often overlooked in assistive technology development. The paper makes a compelling case that access to creative activities is not a luxury but a developmental necessity for children with severe motor impairments, as drawing supports social, emotional, educational, and creative development. For practitioners, the two-state looking/drawing interaction model is a transferable design pattern applicable to any eye-controlled spatial task where the Midas touch problem arises. The CPM-GOMS analysis methodology demonstrates how cognitive engineering can inform accessible interface design by modeling the interplay of perceptual, cognitive, and motor processes. A limitation is that the initial evaluation used only nondisabled children, though the researchers acknowledged this and had a study with disabled users in progress. The project also highlights the importance of involving end users — the power user who provided feedback by eye-typed email exemplifies meaningful participatory design.
Tags: eye tracking · gaze input · motor accessibility · children · creative arts · assistive technology · drawing · dwell time · interaction design