Publications

We propose and demonstrate a linearized model for phase diversity wavefront sensing, facilitating real-time processing and much less data required for training.

We propose a self-supervised method tailored for coded aperture snapshot spectral imaging (CASSI).

We show that significant limitations exist arising from the lack of diversity in the prevailing hyperspectral image datasets.

We propose a comparative analysis of DOEs fabricated by multi-level photolithography and grayscale lithography.

We employ B-spline parameterization for the DOE surface geometry representation for RGBD imaging.

We present DeepLens, a method to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention.

We propose a method to split conventional camera apertures into two halves along with a dual-pixel sensor for various computational imaging applications.

We explore the use of hexagonal grids as a new grid structure for DOE design and fabrication.

We propose an alternative method for fabricating DOEs without RIE lag and with improved surface smoothness.

We propose a new way to fabricate multi-level DOEs by directly growing an optically transparent material on a glass substrate.

We explore bridging the gap between aberrations and depth-from-focus through aberration-aware training (AAT).

We propose a generalized Fourier-based PDWS method by combining an object-independent network with a system-independent image processing procedure.

We propose an end-to-end joint optimization method to learn a DOE placed in front of a projector lens and a compensation network for deblurring.

We propose a lensless camera with an optimized Voronoi-Fresnel phase by maximizing the Modulation Transfer Function volume.

We propose an effective neural network based on low-frequency coefficients in the Fourier domain to determine a better estimate of the unknown aberrations.

We present LLOWFS closed-loop laboratory results under simulated post-Adaptive Optics residuals of GPI 2.0 and simulations of the LLOWFS and FAST sensors for SPIDERS

We present here results from NRC’s NEW-EARTH lab testing of the Fast Atmospheric SCC Technique, a variant of the SCC and its integration with a Lyot-stop Low-Order Wavefront Sensor.

The NRC Canada is funding two projects, the SPIDERS pathfinder at the Subaru telescope (ETA 2023), and the CAL2 upgrade of the Gemini Planet Imager-2 (ETA 2024), to deploy a modified self-coherent camera (based on FAST) to measure the focal plane electric field, and to apply wavefront corrections in a closed-loop down to 10s of ms in a narrow band.

We propose a monocular single-shot imaging approach that optically cloaks obstructions by emulating a large array.

We propose a fully differentiable optical design method for lens design.

We present a thorough investigation of more than 25 state-of-the-art spectral reconstruction methods.

This review reports the latest methods, algorithms, and applications in computational imaging.

We introduce a new dataset synthesis method as well as a novel neural network architecture for single image reflection removal.

We propose an additive lithographic fabrication process to realize simultaneous micrometer and millimeter features.

We propose an etch-free additive lithographic fabriction method for versatile reflective and transmissive DOEs.

We propose a deep unrolling network to examine the problem of illumination spectra estimation in multispectral images.

We propose a general end-to-end complex lens design framework enabled by a differentiable ray tracing image formation model.

We propose a joint design for snapshot HDR imaging with a spatially varying modulation mask.

We propose a polarization demosaicking algorithm for both monochrome and colour DoFP cameras.

We porpose a combinatory method to improve phase retrieval performance.

We present a valid polarization-based reflection contaminated image synthesis method.

We propose a method for snapshot HDR imaging by learning an optical HDR encoding with a DOE in a single image.

We demonstrate a quantitative imaging of phase and bright field amplitude using collimated white light illumination.

We propose an approach for HDR image reconstruction using a spatially-varying mask and an inception network.

We propose a polarization demosaicking algorithm for both monochrome and color DoFP cameras.

We propose a snapshot hyperspectral imaging system that employs a diffractive optical element and an end-to-end network.

We propose a new formula to connect between slopes wavefront sensors and curvature sensors.

We demonstrate an adaptive optics system for regular cameraswith unprecedented capability to sense and correct large distortions.

We propose a system to create high‐quality 1080p live MR footage, enabling realistic virtual experiences to be shared among a number of people.

We propose a reconfigurable rainbow PIV system that extends the volume size to a considerable range.

We describe how scene depth can be extracted using a hyperspectral light field capture (H-LF) system.

We present a single camera hyperspectral light field imaging solution with spectral coded catadioptric mirror arrays.

We introduce a Coded Wavefront Sensor that provides high spatio-temporal resolution using a simple masked sensor, under white light illumination.

We propose a mixed reality system that overlays the virtual world with real world objects captured by a Kinect depth camera.

We propose an improved particle swarm optimization method of ultra-thin diffractive optical elements design for multicolour beam shaping.

We introduce numerically optimized encoded phase masks for focus and zoom through changes in the mechanical alignment.

We introduce a diffractive achromat based on computational optimization and a corresponding algorithm for correction of residual aberrations.

We jointly design lightweight diffractive-refractive optics and post-processing algorithms to enable imaging under white light illumination.