Research
Selected research problems in visual perception and computational imaging.
Monocular Metric Depth Estimation
Estimating depth in metres from a single RGB image, with a focus on the reliability challenges faced by robotic and UAV perception when reference depth is unavailable.
DARPA CIDAR Challenge
Passive image-based distance measurement at kilometer ranges, using multimodal imagery and incomplete information about scene and imaging conditions. Purdue was one of five semifinal awardees.
Pupil Design for Computational Wavefront Estimation
Asymmetric pupil designs for wavefront phase recovery. I designed, built, and refined the optical testbed and collected more than 3,000 experimental PSF images, with 2,500 selected for the final experiments.