Motion deblurring using spatiotemporal phase aperture coding

• DOI: 10.1364/optica.399533
• Publication Date: 2020-08-19T11:30:12Z
• Authors: Shay Elmalem; Raja Giryes; Emanuel Marom
• Keywords: Deblurring; Motion blur

Motion-related image blur is a known issue in photography. In practice, it limits the exposure time while capturing moving objects; thus, achieving proper difficult. Extensive research has been carried out to compensate for it, allow increased light throughput without motion artifacts. this work, joint optical-digital processing method deblurring proposed and demonstrated. Using dynamic phase coding lens aperture during acquisition, trajectory encoded an intermediate optical image. This embeds cues both direction extent by coloring spatial of each object. These color serve as guidance digital process, implemented using convolutional neural network (CNN) trained utilize such restoration. Particularly, unlike previous solutions, our strategy encodes with no limitation on direction, sacrificing efficiency. We demonstrate advantage approach over blind methods coding, well other solutions that use coded simulation real-world experiments.