Understanding context.
Shaping interaction.

I connect sensing, interpretation, and interaction to build systems people can understand and control.

Across mobile, wearable, and spatial interfaces, context includes people’s goals—not only sensor observations.

Three connected questions

From the physical world to the person.

DESIGN ACROSS THE LOOP

Privacy · Safety · Trustworthiness

Three connected capabilities within a shared design boundarySense, Reason, and Adapt overlap around context-aware AI. Privacy, safety, and trustworthiness form a dashed boundary around all three. People retain opportunities to understand and correct responses. The diagram maps a research agenda, not the capabilities of every individual system.SENSEREASONADAPTContext-awareAI
Connected capabilities, shaped by human feedback and shared design responsibilities.

SENSE

Acquire the relevant detail.

Capture useful detail through visual, wearable, and multimodal sensing.

MicroCam ↗RadarFoot ↗

Selected perspectives on the work

One loop, many entry points.

Not every project implements a full loop. The map separates primary, supporting, and unevaluated capabilities. Design Lens records a cross-cutting concern—not a fourth stage or a trustworthiness score.

Primary contributionSupporting componentNot evaluatedReviewed in the survey

  1. MicroCam

    Vision-based surface sensing

    SENSE
    Primary
    REASON
    Supporting
    ADAPT
    Not evaluated

    DESIGN LENS

    Keep claims within the captured surface context.

    Microscope-camera sensing and surface classification. Adaptive applications remain a direction, not an evaluated loop.

  2. RadarFoot

    Wearable radar sensing

    SENSE
    Primary
    REASON
    Supporting
    ADAPT
    Not evaluated

    DESIGN LENS

    Separate recognition performance from downstream usefulness.

    Ground-surface recognition with foot-worn radar; downstream interface usefulness was not evaluated.

  3. MultiEEG-GPT & IntervEEG-LLM

    Physiological and contextual interpretation

    SENSE
    Supporting
    REASON
    Primary
    ADAPT
    Not evaluated

    DESIGN LENS

    Distinguish physiological interpretation from diagnosis.

    Related but distinct EEG–context studies; neither establishes clinical diagnosis or an autonomous intervention loop.

  4. Ambient2Hap

    Accepted, in press · ICME 2026

    SENSE
    Supporting
    REASON
    Supporting
    ADAPT
    Primary

    DESIGN LENS

    Consider comfort, calibration, and the feedback context.

    Visual and auditory cues become body-based haptic feedback—not a new tactile sensor.

  5. Speech2Pro

    Revise & resubmit · IJHCS

    SENSE
    Supporting
    REASON
    Primary
    ADAPT
    Primary

    DESIGN LENS

    Keep authoring decisions open to user correction.

    Speech and spatial context guide projection authoring. Wizard-of-Oz alignment is not fully autonomous operation.

  6. SituGuard

    Context-sensitive privacy · Interactive prototype

    SENSE
    Supporting
    REASON
    Supporting
    ADAPT
    Supporting

    DESIGN LENS

    Primary focus: privacy choices and perceived user control.

    Context-sensitive privacy choices in a 12-participant interactive-mockup study—not deployment privacy or safety certification.

  7. VMIs

    Vision-based multimodal interfaces · Survey of 109 papers

    SENSE
    Reviewed
    REASON
    Reviewed
    ADAPT
    Reviewed

    DESIGN LENS

    Primary focus: a design taxonomy and cross-cutting challenges.

    A taxonomy of 109 papers. Reviewed marks literature coverage, not an implemented or evaluated capability.

Research outputs

Selected publications.

Further publications; system case studies appear in the Portfolio.

2026

Speech2Pro: Using Speech-Driven Image Creation for Enhancing Stage Performance Projection

International Journal of Human–Computer Studies

Revise and resubmit (R&R) · First author

2026

Ambient2Hap: Rendering Haptic Feedback via Somatosensory Electrical Stimulation from Ambient Cues in Virtual Reality

IEEE International Conference on Multimedia and Expo (ICME)

Accepted, in press · First author

2026

SpeechAgent: An End-to-End Mobile Infrastructure for Speech Impairment Assistance

Interspeech

Co-author

2026

EmoDrink: An Embodied MR Framework for Physiology-Informed Wellbeing Recommendations, Illustrated Through Beverage Choice

Augmented Humans International Conference (AHs)

Co-author

2025

Towards Human–AI Synergy in UI Design: Supporting Iterative Generation with LLMs

ACM Transactions on Computer-Human Interaction (TOCHI)

Co-author

2025

Toward AI-driven UI Transition Intuitiveness Inspection for Smartphone Apps

International Journal of Human–Computer Studies

Co-author

2025

A Fine-grained Context Awareness with Radar-based and Vision-based Surface Sensing

Radar-Based Human–Computer Interaction, Springer Nature

Book chapter, accepted in press · First author