I research in the areas of virtual reality, computer graphics and machine vision.
My focus is filmed virtual reality experiences, like 360˚ and volumetric videos.
Producing volumetric video can be expensive due to the amount of time it takes to process the 3D meshes. This project, done partly while I was a Researcher in Residence at Digital Catapult, looks to reduce this cost using 3D skeleton tracking. We use these skeletons before 3D reconstruction to identify ways to loop and sequence clips of volumetric video, meaning fewer frames need to be processed. Our hope is that this technique can also be used to aid real-time, on-set decision making, further reducing costs incurred from re-shoots. I presented a poster on preliminary findings from this project at IEEEVR 2020 (open access).
In volumetric video, eye gaze presents many issues. If a filmed character is looking at the user and the user moves, for example, the best way to ensure eye lines remain correct is an open question. To address these issues, one of the fundamental aspects we must understand is how accurately users can assess where a volumetric character is directing their gaze. At IEEEVR 2019 I presented the paper Perception of Volumetric Characters' Eye-Gaze Direction in Head-Mounted Displays (open access), written with Prof Anthony Steed. This paper investigates how accurately people can gauge where a volumetric avatar is looking.
At IEEEVR 2018 I presented the TVCG paper The Effect of Transition Type in Multi-View 360-Degree Media, written with Prof Anthony Steed. In this paper, we looked at the impact of the transition type a user sees while navigating a scene captured as multi-view 360-degree media. Aspects measured included spatial awareness, preference, and the subjective feeling of moving through the space.
At IEEEVR 2017 I presented the paper Cinematic virtual reality: Evaluating the effect of display type on the viewing experience for panoramic video, written with Prof Anthony Steed. Several metrics were explored that may indicate advantages and disadvantages of cinematic virtual reality compared to traditional viewing formats. We explored the consumption of panoramic videos in three different display systems: a HMD, a SurroundVideo+ (pictured), and a standard 16:9 TV. Aspects examined were spatial awareness, narrative engagement, enjoyment, memory, fear, attention, and a viewer’s concern about missing something.
As part of my work with panoramic video, I presented a paper investigating object removal in 360° media - written with Prof Anthony Steed - at the Conference on Visual Media Production (CVMP). The work also included helping out on a couple of 360° film shoots for BBC R&D.
A project to create an interactive fog screen. Atomised water was used to create a plane of fog. Back projection was used to create visuals that appear to float in mid-air. A leap motion was added to allow the user to reach in and interact with these objects. Due to the limited viewing angle caused by the mist's light scattering, we were also able to use the fog screen as a parallax barrier to create a multi-viewpoint 3D display. Details on construction and results are available on the project's blog.
A group of us from UCL created a VR escape the room puzzle game. The game was originally built for UCL's CAVE using the Unity game engine, as shown in this video. I recently ported it to work in the HTC VIVE.
As a teaching assistant I run tutorials, labs and demos on several topics at UCL. On the Virtual Environments module, I run classes on Unity, 3D modelling software and VR hardware. I've also tutored in computer graphics, robotics, systems design and various kinds of programming.