Shuffled Rolling Shutter Camera
Published in Springer International Publishing, 2023
Recommended citation: E. Vera, F. Guzmán and N. Diaz, “Shuffled Rolling Shutter Camera,” Springer International Publishing, pp. 499-513, 2023. [DOI].
This chapter provides a comprehensive and up-to-date review of the rolling shutter (RS) mechanism found in complementary metal-oxide semiconductor (CMOS) imaging sensors as a coding opportunity for high-speed videos. Based on this, we present a cutting-edge readout architecture called shuffled rolling shutter (SRS) that optimally designs the scanlines entries of the RS for improved sampling of the space-time datacube. The RS restriction implies that only one entry per column is sampled at each exposure time. This restriction can be modeled using a three-dimensional N2 queens problem (3DN2QP), such that the optimal design of entries is computed straightforwardly, solving a sphere packing (SP) problem. The main advantage of packing equal-size spheres in a cubic container is promoting uniform spatiotemporal sampling. Simulation and experimental results prove that with the captured measurements using the proposed SRS, we can recover the underlying video from a snapshot. This opens the avenue for the development of new CMOS imaging detectors with shuffled positions of the RS, now able to turn a typical nuisance into an opportunity to capture and recover videos from snapshots.
Cite
@Inbook{Vera2024,
author="Vera, Esteban
and Guzman, Felipe
and Diaz, Nelson",
editor="Liang, Jinyang",
title="Shuffled Rolling Shutter Camera",
bookTitle="Coded Optical Imaging",
year="2024",
publisher="Springer International Publishing",
address="Cham",
pages="499--513",
abstract="This chapter provides a comprehensive and up-to-date review of the rolling shutter (RS) mechanism found in complementary metal-oxide semiconductor (CMOS) imaging sensors as a coding opportunity for high-speed videos. Based on this, we present a cutting-edge readout architecture called shuffled rolling shutter (SRS) that optimally designs the scanlines entries of the RS for improved sampling of the space-time datacube. The RS restriction implies that only one entry per column is sampled at each exposure time. This restriction can be modeled using a three-dimensional N2 queens problem (3DN2QP), such that the optimal design of entries is computed straightforwardly, solving a sphere packing (SP) problem. The main advantage of packing equal-size spheres in a cubic container is promoting uniform spatiotemporal sampling. Simulation and experimental results prove that with the captured measurements using the proposed SRS, we can recover the underlying video from a snapshot. This opens the avenue for the development of new CMOS imaging detectors with shuffled positions of the RS, now able to turn a typical nuisance into an opportunity to capture and recover videos from snapshots.",
isbn="978-3-031-39062-3",
doi="10.1007/978-3-031-39062-3_27",
url="https://doi.org/10.1007/978-3-031-39062-3_27"
}