Abstract
We study encoded diffraction with shallow neural networks using singular value decomposition entropy (HSVD) to measure image span. Higher-HSVD synthetic training images are learned more slowly; their generalized models generally attain higher reconstruction accuracy.
© 2022 The Author(s)
PDF Article | Presentation VideoMore Like This
Md Sadman Sakib Rahman and Aydogan Ozcan
FM2H.7 CLEO: QELS_Fundamental Science (CLEO:FS) 2022
Baurzhan Muminov, Altai Perry, Rakib Hyder, M. Salman Asif, and Luat T. Vuong
ITh5D.5 Imaging Systems and Applications (IS) 2021
Feng Tian and Weijian Yang
JW3A.7 CLEO: Applications and Technology (CLEO:A&T) 2022