PerEye: Co-Designing Extended Reality Rendering Attributes for Vision Health Diagnosis and Education
PerEye is compelling because it reframes XR for vision health around asymmetric per-eye rendering instead of the default assumption of identical binocular input. The paper’s contribution is strongest as a co-designed translational artifact: it defines clinically meaningful rendering attributes, shows a feasible but non-interchangeable diagnostic probe, and demonstrates measurable empathy gains in education.
Video Figure
Axes Lens
Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.
Contribution shape
- Knowledge form
- technical knowledge typical · 50/268
- Novelty type
- artifact typical · 20/268
- Abstraction level
- artifact typical · 19/268
- Generalization target
- task class typical · 63/268
- Validation mode
- mixed methods typical · 136/268
Evidence profile
- Evidence strength
- strong typical · 158/268
- Claim alignment
- strong typical · 231/268
- Overclaim risk
- low typical · 53/268
Review Summary
PerEye stands out because it does more than present another XR prototype for healthcare. The paper identifies a basic but often unexamined assumption in mainstream XR design: that both eyes should receive symmetrical, matched rendering. In vision health, that assumption is often wrong, because many clinically meaningful conditions are asymmetric, dynamic, or otherwise poorly represented by standard binocular pipelines. The introduction makes that departure explicit, and the paper’s contribution is to operationalize an alternative through PerEye, a co-designed research probe that enables gaze-contingent control of visual attributes for each eye. What makes the paper especially strong for CHI is the way it connects artifact design, domain expertise, and empirical validation across multiple contexts. Rather than claiming a universal clinical solution, the authors build a trajectory: clinicians help define the rendering attributes, those attributes are embedded into a VR visual field assessment, and the same design logic is then used in an educational simulation for orthoptic training. That gives the work coherence across studies and makes the artifact legible as a reusable design approach rather than a one-off demo. The evidence is also appropriately scoped. Study 2 is encouraging but clearly bounded: the VR system broadly captures relative sensitivity trends, yet the paper explicitly reports modest correlations, proportional bias, and variability that constrains interchangeability with the Humphrey Field Analyzer. That is exactly the kind of result that should be read as feasibility evidence, not as proof of clinical replacement. Study 3 is stronger on its own terms, showing statistically significant empathy gains with a large effect size, but it is still based on a small, specific learner population. Taken together, the paper succeeds because it does not overclaim. Its real significance is in showing that clinician-informed, asymmetric per-eye rendering is a meaningful XR design space for diagnosis, communication, and education in vision health. That combination of conceptual reframing, concrete artifact work, and honest empirical boundaries is what makes it a high-quality CHI contribution.
What Changed
Canon before
Most XR systems assume symmetrical, static, and standard binocular vision with identical rendering to both eyes, based on common-sense assumptions of matched binocular input. Clinical vision impairments, however, frequently involve well-documented asymmetric and dynamic visual conditions that XR fails to represent accurately. Prior approaches have been driven more by technological capability than by co-design with vision domain experts. XR devices also typically offer limited eye-tracking frequency and proprietary data access that constrain clinical precision and integration.
Departure from common sense
The paper breaks the common-sense assumption that both eyes' visual inputs and rendering in XR should be symmetrical and identical. It shows that asymmetric, monocular rendering with fine-grained control per eye is crucial to realistically replicate clinical conditions and enhance diagnostic and educational applications in vision health.
Actual novelty
This work introduces PerEye, a co-designed XR research probe enabling independent, gaze-contingent rendering attribute control for each eye in an HMD. It establishes clinician-informed rendering parameters mapped to vision health needs, implements a VR-based visual field assessment prototype demonstrating diagnostic feasibility, and applies immersive vision simulations in orthoptic training to improve empathy and understanding. It advances a sustained multi-study co-design trajectory aligning XR rendering with clinical reasoning and education in vision health.
Evidence
The paper supports its claims with three linked studies visible in the provided sections. The introduction states the core departure from symmetric binocular rendering assumptions and introduces PerEye as a co-designed research probe with per-eye gaze-contingent control. Study 2 reports that the VR visual field prototype captures relative sensitivity patterns but shows only modest correlation and clear limits on interchangeability with Humphrey perimetry. Study 3 reports statistically significant empathy gains after simulation in orthoptic students. Overall, the evidence is strong for a scoped translational artifact contribution, not for clinical replacement.
“Our work addresses these gaps by introducing PerEye, a co-designed XR research probe that enables gaze-contingent control of visual attributes for each eye in head-mounted displays. Figure 2 summarises our multi-year trajectory, in which we worked with vision clinicians to move from defining rendering attributes, to prototyping a diagnostic proof of concept, to educational application”
actual novelty · 1 Introduction · confidence 0.97
“ Our paper asserts that most XR systems, including those used in vision health, rely on rendering assumptions rooted in symmetrical, static, and standard binocular vision, where both eyes are expected to converge and form aligned visual input [53]. This assumption contradicts well-documented ophthalmic asymmetries in visual function [36] and limits”
departure from common sense · 1 Introduction · confidence 0.95
“Analysis revealed that the VR prototype broadly captured relative sensitivity trends but did not achieve strong absolute agreement with the HFA benchmark. As shown in Figure 8, Bland–Altman analysis indicated a small positive bias of 0.065 (VR minus HFA), with LoA ranging from − 0.329 to 0.459, reflecting variability that constrains interchangeability between the two methods. To fur”
limitation · 5.4.2 Agreement Between VR and HFA Sensitivity Values. · confidence 0.93
“ble dispersion. Taken together, the Bland–Altman findings, proportional bias regression, and correlation results position the VR system as a feasible probe for tracking relative patterns in visual sensitivity rather than a clinically interchangeable replacement for gold standard perimetry”
validation scope · 5.4.2 Agreement Between VR and HFA Sensitivity Values. · confidence 0.92
Limits
Method limits
The strongest explicit limitation in the provided sections is in Study 2: the VR prototype did not achieve strong absolute agreement with the HFA benchmark, showed variability that constrains interchangeability, and systematically underestimated low sensitivities where luminance capabilities differ. Study 3 also uses a small orthoptic-student sample, which narrows generalization.
Deployment limits
Clinical deployment is limited because the prototype is positioned as a feasible probe rather than a clinically interchangeable replacement for gold standard perimetry. Its performance depends on overlap between VR and HFA luminance capabilities, and the educational findings come from a specific training context rather than routine clinical deployment.
Boundary conditions
The contribution is best read as applying to XR vision-health tasks where asymmetric per-eye rendering matters, especially portable probing of visual sensitivity patterns and educational simulation. The diagnostic prototype is bounded to relative pattern tracking rather than absolute clinical equivalence, and the empathy findings are bounded to orthoptic training participants.
Position in field
This paper sits at the intersection of HCI, XR, and vision health. Its main contribution is not a validated medical device but a co-designed translational artifact and research trajectory that makes asymmetric per-eye rendering actionable across diagnosis and education. Within CHI, it is strongest as an artifact-centered, clinically grounded systems contribution with appropriately scoped empirical validation.