Smart Glasses for CVI: Co-Designing Extended Reality Solutions to Support Environmental Perception by People with Cerebral Visual Impairment
Bhanuka Gamage, Nicola McDowell, Dijana Kovacic, Leona Holloway, Thanh-Toan Do, Arthur James Lowery, Nicholas Price, Kim Marriott · 2025 · ASSETS 2025: 27th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663547.3746383
Summary
This paper presents the first co-design study developing and evaluating smart glasses as an assistive platform for adults with cerebral visual impairment (CVI). CVI is set to become the leading cause of vision impairment, yet it remains dramatically underrepresented in assistive technology research. Unlike ocular vision impairment that stems from damage to the eyes, CVI arises from disruptions to the brain's visual processing centres, affecting higher-order visual functions such as object recognition, facial perception, visual attention, and the ability to process multiple visual elements simultaneously (simultanagnosia). The researchers conducted an eight-month co-design study with two adults with CVI — both of whom are co-authors — using the Double Diamond design framework. The study comprised three phases: a two-week diary study (Discover phase) where co-designers documented daily CVI-related challenges across four selected days; two ideation and scoping workshops (Define phase) where difficulties were articulated and potential smart glass solutions brainstormed; and ten iterative development workshops (Develop/Deliver phases) using the Apple Vision Pro to prototype, test, and refine solutions. Six key difficulties were selected for development: locating items in cluttered environments, reading text, face-to-face conversations, recognising people, visual attention, and stress from sensory overload. The study explored over 80 distinct design options across these six challenges, iteratively refining solutions based on direct co-designer feedback in their home environments.
Key findings
Smart glasses effectively addressed four of six challenges. For locating objects, co-designers preferred stable visual highlights (outlines in personalised colours) with environment darkening for contrast, plus large directional arrows for off-screen objects — auditory cues were rejected as unreliable. For reading text, a combined text-to-speech with visual overlay approach was strongly preferred, along with a "subtle scaling" feature that slightly enlarges text lines as the user's gaze passes over them. For face-to-face conversations, a fixed rectangular frame around the speaker's upper body with dimmed surroundings dramatically improved visual integration — one co-designer described it as recreating what vision would feel like without CVI. For recognising people, virtual name tags positioned below the chin (not above the head, due to visual field issues) with gaze-responsive scaling and fading were effective. Visual attention features showed promise as training tools but were not effective for everyday real-time use — attention-guiding markers caused sensory overload and discomfort. For sensory overload, immersive virtual environments provided "calm spaces" for sensory breaks, with adjustable reduction modes (minimal, medium, extreme) and potential for biometric integration via heart rate variability monitoring. Critical design considerations included hyper-personalisation (even two co-designers had markedly different preferences due to distinct CVI profiles), context-aware adaptation, and minimising cognitive load by avoiding moving, flashing, or dense visual augmentations.
Relevance
This research fills a significant gap in assistive technology for CVI — a condition affecting increasing numbers of people yet largely ignored by AT researchers who have focused primarily on ocular vision impairment. The findings have broad implications for accessibility practitioners. First, CVI challenges existing assumptions about visual accessibility: because CVI affects higher-order processing rather than acuity, traditional accommodations like magnification may be insufficient or even counterproductive. Second, the co-design methodology demonstrates how deeply personalised assistive solutions must be — even within the same diagnosis, the two co-designers had substantially different needs and preferences, underscoring the limitations of one-size-fits-all approaches. Third, the study reveals that smart glasses can serve a dual role as both real-time assistive tools and rehabilitation aids that help users develop compensatory visual strategies over time. For web and digital accessibility, the findings about sensory overload, visual clutter, and the need for environment simplification parallel challenges in designing accessible digital interfaces for people with cognitive and visual processing differences.
Tags: cerebral visual impairment · CVI · smart glasses · extended reality · augmented reality · co-design · Apple Vision Pro · assistive technology · visual processing · sensory overload · simultanagnosia