Computer vision typically regards images as given entities to be processed.
However, richer information can be extracted by modifying and analyzing
the imaging process itself. This modification includes the sensor or the
illumination, in conjunction to carefully tailored algorithms. We term this
approach “hybrid imaging”, since it exploits the advantages of both the
sensor and the algorithmic components of a vision system, treating them
as a whole.
In this talk, we describe our recent results in this context. We show
a bio-inspired method for overcoming underwater visibility problems. We
then describe a method for blindly estimating simultaneous spatio-temporal
inconsistencies of sensors (gain, vignetting, radiometric response). In
addition, we handle needs of imaging objects under various illumination
settings. We address these needs with greatly improved quality using
multiplexing of light sources, followed by computational decoupling.
Finally, we go beyond optics to explore audio-visual interaction, whereby
a novel computer vision algorithm uniquely pinpoints the pixels that
correspond to sound sources, with high accuracy.