Neil Fan Email & Phone Number

Project: TangiAR: Markerless Tangible System for Augmented Reality with Everyday Objects (first author)Augmented Reality (AR) bridges the gap between the real and digital worlds. However, conventional control methods like gestures and controllers often overlook the integration with the real-world environment. In this project, my goal was to create a system that empowers users to control it using everyday objects, enhancing the user experience by grounding digital interactions in the physical world. Furthermore, I optimized the system's performance to ensure it operates efficiently on smart glasses alone, making advanced AR technology more accessible and user-friendly.* Developed the first interaction system in the literature that operates in real-time on head-mounted devices (e.g., Microsoft HoloLens 2), using everyday objects to control the system, such as naturally using an eraser as a character in an AR game. * Tracking objects via a ported iterative optimization-based computer vision algorithms, converting their 6 Degrees of Freedom poses into control signals for AR applications* Supported different touch interactions with various fingers on objects. * Packed the system as a C++ plugin for Unity applications, allowing other developers to use it easily.* Utility tests showed significant improvements in both task completion time and user workload. * Authored a paper currently under submission.Video Link: https://youtu.be/uVoKkggZEXI

Vancouver, British Columbia, Canada

Project: Touch Tracking and Latency Reducing on Ad Hoc Surfaces (first author)If touch sensing is not limited to smartphones or tablets, we can map computer interfaces onto daily surfaces, such as tables or walls, in AR. This turns them into large touch screens. However, detecting touch on surfaces without sensors is challenging. I conducted an independent research project where I:* Proposed a novel machine learning pipeline to enhance touch identification accuracy on surfaces without sensors, achieving state-of-the-art performance (4% improvement over previous benchmarks).* Used a Computer Vision algorithm with depth camera inputs for touch detection.* Reduced touchdown latency from 230ms to 70ms, with a minimal accuracy trade-off.* Compiled an uncompressed specialized dataset (1.6TB) designed explicitly for touch.* Paper Accepted and presented at ACM Interactive Surfaces and Spaces (ISS, 2022)

Honda ASIMO Humanoid Robotics Artificial Skin ProjectSupervisor: Dr. John MaddenThis project aims to design a flexible pressure and shear sensor for humanoid robots to mimic human skins. Initially joining the project as a research intern, my dedication and contributions quickly led to my promotion to team leader within eight months.As the leader of a diverse team comprising seven undergraduate and graduate students, I played a pivotal role in steering the project's direction. My responsibilities expanded to include problem identification and providing unofficial supervision of the research efforts of team members. During this time, I:* Designed and implemented a sensor architecture that reduced the complexity of a previous iteration by multiplexing and repurposing an electrode to act both as a ground shield and a self-capacitive sensing terminal.* Redesigned the hardware and optimized the code to reduce the cross-talk by an order of magnitude.* Experimented and optimized electrode material conductivity and improved the product manufacturing success rate by over 100%.* Collected distinguishable pressure and proximity data for machine learning data training.

Master Of Science - Ms, Computer Science, 4.10/4.30

Research Supervisor: Dr. Robert Xiao

Bachelor Of Applied Science - Basc, Electrical And Electronics Engineering, 4.10/4.30