2026 · Research
Spatial-IQ
Hierarchical capability tests help pinpoint where multimodal models succeed or fail at spatial reasoning.
Projects
Spatial intelligence, immersive systems, and technologies that connect human vision with computation.
Recent work
2026 · Research
Radiance fields on light field displays
Efficient rendering connects radiance field representations with interactive, glasses-free 3D displays.
2026 · Research
Natural walking in immersive VR
Studies of hemianopia and cerebral visual impairment measure pedestrian detection, avoidance, and visual scanning during natural walking.
2025 · VSS
Perceptual evaluation of video artifacts
Human-observer experiments examine artifact detection across different video comparison interfaces.
Jung Lab research · 2015–2024
Research programs developed at Harvard Medical School and Mass Eye and Ear.
Virtual Reality Simulation of Mobility
Immersive VR walking environments measure how people detect and avoid approaching pedestrians. Built with Unity and head-mounted displays, the simulations reproduce crowded spaces while recording natural walking behavior.
Funding & support
- 09/2020 – 08/2024, Monocular Visual Confusion for Field Expansion, NIH/NEI R01 – EY031777
Augmented Reality for Field Expansion (Vision Multiplexing)
Optical see-through devices and multiplexing prisms expand the visual field for people with field loss. This work studies how visual confusion, binocular rivalry, contrast, depth cues, and motion affect mobility.
Field Expansion for Acquired Monocular Vision Using a Multiplexing Prism The effect of visual rivalry in peripheral head-mounted displays on mobility
Funding & support
- 09/2020 – 08/2024, Monocular Visual Confusion for Field Expansion, NIH/NEI R01 – EY031777
- 07/2018 – 12/2019, Field Expansion for Acquired Monocular Vision using Multiplexing Prism, Fight for Sight Grant #GA18003
Computational Model of Visual Perception and Binocular Vision
Computational and psychophysical studies examine how overlapping views are combined or compete in binocular vision. The goal is to understand visual perception in field-expansion devices and AR displays.
Develop then Rival: A Human Vision-Inspired Framework for Superimposed Image Decomposition Binocular double vision in the presence of visual field loss
Funding & support
- 09/2020 – 08/2024, Monocular Visual Confusion for Field Expansion, NIH/NEI R01 – EY031777
- 09/2019 – 08/2024, Visual field expansion through innovative multi-periscopic prism design, NIH/NEI R01- EY023385 (Co-Investigator)
Active Confocal Imaging for Visual Prostheses
Light-field imaging emphasizes objects of interest while suppressing background clutter. Active confocal imaging explores front-end processing for retinal and cortical implants and sensory substitution devices.
Active Confocal Imaging for Visual Prostheses Comparing object recognition from binary and bipolar edge images for visual prostheses
Funding & support
- 01/2016 – 01/2021, Active confocal imaging for visual prostheses, US Department of Defense, W81XWH-16-1-0033
- 02/2015-10/2015, Active confocal imaging for visual prostheses, Grant from the Promobilia Foundation, Sweden #14222
Light-Field Imaging and Computational Photography
Integral imaging uses microlens arrays to capture and reconstruct 3D scenes. The work includes real-time reconstruction, depth extraction, confocal image generation, and background suppression for object recognition.
Real-time integral imaging system for light field microscopy Real-time capturing and 3D visualization method based on integral imaging
Funding & support
- 01/2016 – 01/2021, Active confocal imaging for visual prostheses, US Department of Defense, W81XWH-16-1-0033
- 12/2012-12/2013, Study on effect of super multi-view condition in three-dimensional display to accommodation response and improvement of optical vision rehabilitation device, National Research Foundation of Korea, 2012R1A6A3A03038820 (PI)