Overview

liVeR

Enhancing liver surgical planning and education at Canada's top hospital by bringing 3D to life.

This was a self-led and funded Master's capstone project where I led collaboration with clients and spearheaded the end-to-end design process. Besides design, I also scripted in Unity to make functional, complete VR applications. In the end, these well-received apps incentivized continued financial support and are poised to augment the workflows of the liver team at the Toronto General.

Role

Product Designer, Unity Developer

Duration

16 months (part-time)

Outcome

2 VR, 1 desktop apps (functional)

Tools

Figma, Illustrator, Photoshop, After Effects, Premiere Pro, Blender, Unity, Meta VR SDKs, Normcore, GitHub, Visual Studio

Figma, Illustrator, Photoshop, After Effects, Premiere Pro, Blender, Unity, Meta VR SDKs, Normcore, GitHub, VS

Clients

Toronto General Hospital:
Dr. Chaya Shwaartz (Director, HPB)
Albert Fung (Team Lead, TVASurg)

Toronto General Hospital:
Dr. C. Shwaartz (Director, HPB)
Albert Fung (Team Lead, TVASurg)

Toronto General Hospital:
Dr. C. Shwaartz
(Director, HPB)
Albert Fung
(Team Lead, TVASurg)

Advisory Team

Albert Fung (Unity engineer)
Dr. Chaya Shwaartz (product)
Dr. Derek Fung (product)
Dr. Jodie Jenkinson (research)
Dr. Prachi Patel (research)

Project

Self-led, funded Master's capstone

1.

The project produced a multi-user application for internal use.

The flow enables users to easily find, review, and discuss patient-specific anatomy.

2.

We also built twin desktop and VR tools to be rigorously compared in a scientific study.

The applications carefully onboard users to the technology and then allows the user to review patient-specific anatomy.

Research

Tasked with an early vision, I did research, including 4 interviews, to pinpoint core user pain points, enterprise needs, and technical constraints.

In short, liver anatomy interpretation is a time-consuming, challenging process ripe for innovation.

Techniques like personas, user journey maps, and affinity diagrams helped me distill the data into project requirements.

As an analogy, surgeons and trainees rely on the left (sparse cross-sections) to mentally reconstruct the right (patient-specific anatomy).

As an analogy, surgeons and trainees rely on the top (sparse cross-sections) to mentally reconstruct the bottom (patient-specific anatomy).

In short, liver anatomy interpretation is a time-consuming, challenging process ripe for innovation.

Roadmap

What if surgeons and trainees could halve their anatomy interpretation time for every single patient?

I outlined our project requirements as follows: one multi-user application for internal use, two twin applications for a scientific study, and one video for VR device donning.

Strategizing

In addition to roadmapping and managing the project, my research also significantly informed platform strategy. I highlighted desktop, to cross-functional stakeholders and clients, as an equally viable modality backed by research.

Iterative Design

Each of the 3 apps underwent 1-3 key iterations. I led cross-functional design reviews and usability testing — involving a total of 10 end users and 5 stakeholders — that informed each cycle.

Tests and reviews often included semi-structured exploration as well as task-specific aspects, focusing on particular elements like onboarding flows.

Improving User Flows

Improving User Flows

Improving User Flows

Though competitive analysis showed a trend of indirect interactions (i.e. lasers) in pre-game lobbies, user testing encouraged us to rethink the lobby flow. Ultimately, we maximized direct interactions (e.g. grabbing) and increased selection efficiency.

Though competitive analysis showed a trend of indirect interactions (i.e. lasers) in pre-game lobbies, user testing encouraged us to rethink the lobby flow. Ultimately, we maximized direct interactions (e.g. grabbing) and increased selection efficiency.

Though competitive analysis showed a trend of indirect interactions (i.e. lasers) in pre-game lobbies, user testing encouraged us to rethink the lobby flow. Ultimately, we maximized direct interactions (e.g. grabbing) and increased selection efficiency.

Reducing Interaction Cost

Reducing Interaction Cost

Reducing Interaction Cost

Through testing, I iteratively improved this control panel (↑):

  • Removed hard-to-reach buttons

  • Reduced vertical head movement (causes motion sickness)

  • Maintained patterns familiar to users (i.e. Windows' top-right session controls)

  • Better match the twin app’s UI patterns

Through testing, I iteratively improved this control panel (↑):

  • Removed hard-to-reach buttons

  • Reduced vertical head movement (causes motion sickness)

  • Maintained patterns familiar to users (i.e. Windows' top-right session controls)

  • Better match the twin app’s UI patterns

Through testing, I iteratively improved this control panel (↑):

  • Removed hard-to-reach buttons

  • Reduced vertical head movement (causes motion sickness)

  • Maintained patterns familiar to users (i.e. Windows' top-right session controls)

  • Better match the twin app’s UI patterns

Balancing Design with Feasibility

Balancing Design with Feasibility

Balancing Design with Feasibility

We explored object momentum to enhance interaction efficiency and immersion, but we decided to remove it as it interfered with the development of more critical features.

We explored object momentum to enhance interaction efficiency and immersion, but we decided to remove it as it interfered with the development of more critical features.

We explored object momentum to enhance interaction efficiency and immersion, but we decided to remove it as it interfered with the development of more critical features.

What Worked Well For Users

Calling back to the user, enterprise needs and the constraints we identified, we designed for intuitiveness, inclusivity, comfort, and immersion — and limited the scope to ensure a seamless first experience that encourages adoption.

Simplifying the Age-Old Problem (→)

Simplifying the Age-Old Problem

Simplifying the Age-Old Problem

While 2D cross-sections (i.e. CTs) are hard to read, they provide valuable accuracy and detail. To make them simpler, we featured a 3D model, introduced CT color coding, and added indicator planes that correlate 2D and 3D views.

Enterprise Specificity from Co-Creation

I co-created with my clients to reduce user friction as much as possible. Notably I fully implemented the UI shown here to enable high-fi testing.

Visual Onboarding to Reach Wider Demographic (→)

Visual Onboarding to Reach Wider Demographic

Visual Onboarding to Reach Wider Demographic

I embedded many GIFs like this one to onboard users to all 3 applications. As compared to text-heavy instructions, usability tests showed they made onboarding seamless.

Multimodality for Interaction Feedback and Immersion

Audio cues like this one proved to be incredibly important to users.

VR FPS Optimization for Comfort (→)

VR FPS Optimization for Comfort

VR FPS Optimization for Comfort

My multifaceted approach included an in-game FPS test during development, lightweight 3D objects, minimal UI transparency, and reducing the need for head movement.

(←) Leveraging Existing Habits

Leveraging Existing Habits

Leveraging Existing Habits

Contextual inquiries revealed interactions that are familiar to our target users — for 3D navigation, CT scroll, zoom, and pan — and we incorporated these across the 3 apps.

(←) Enterprise Specificity from Co-Creation

(↙) Enterprise Specificity from Co-Creation

I co-created with my clients to reduce user friction as much as possible. Notably I fully implemented the UI shown here to enable high-fi testing.

(←) Multimodality for Interaction Feedback and Immersion

Audio cues like this one proved to be incredibly important to users.

Leveraging Existing Habits (→)

Contextual inquiries revealed interactions that are familiar to our target users — for 3D navigation, CT scroll, zoom, and pan — and we incorporated these across the 3 apps.

(←) VR FPS Optimization for Comfort

My multifaceted approach included an in-game FPS test during development, lightweight 3D objects, minimal UI transparency, and reducing the need for head movement.

What Worked Well For Users

Calling back to the user, enterprise needs and the constraints we identified, we designed for intuitiveness, inclusivity, comfort, and immersion — and limited the scope to ensure a seamless first experience that encourages adoption.

Impact

"You did such a great job, and you really brought a long-awaited vision to life." "I can see myself using this regularly."

liVeR, a brand new product and one of the first of its kind, was tested by stakeholders, staff surgeons, surgical fellows, residents, and medical students. Recognized for its success and potential, stakeholders secured $75,000 in continued grant funding — a significant feat in a nonprofit context. Demonstrating a seamless, time-saving UX, liVeR is now driving the formation of a new team for formal launch preparation.

Metrics

For the internal tool, I would analyze data like session length, error rates, qualitative user feedback, scene/page exit rates or funnelling statistics to gauge success and identify where the experience can be improved.

Reflections

Thanks to this project, I learned the value of proactively thinking at the level of the organization and its long-term objectives. Also, I learned the importance of clear verbal communication as a skill in design, whether that'd be for the purposes of aligning stakeholders on roadmaps, co-creating with domain experts, or getting team buy-in for research initiatives.

Thank you to the team, clients, and all who made this project possible.

Designed and built by Remi (Ling Yan) Gao | 2025 | It's great to see you here!

Designed and built by Remi (Ling Yan) Gao | 2025 | It's great to see you here!

Designed and built by Remi (Ling Yan) Gao | 2025 | It's great to see you here!

Designed and built by Remi (Ling Yan) Gao | 2025 | It's great to see you here!

Designed and built by Remi (Ling Yan) Gao | 2025
It's great to see you here!