Skip to main content
. Author manuscript; available in PMC: 2021 Jan 1.
Published in final edited form as: IEEE J Sel Top Quantum Electron. 2019 Jun 6;26(1):3700114. doi: 10.1109/JSTQE.2019.2921376

Fig. 5.

Fig. 5.

D2NN-based hybrid neural networks. (a) The architecture of a hybrid (optical and electronic) classifier. (b) Final design of phase-only optical layers (Δz = 40×λ) at the front-end of a hybrid handwritten digit classifier with a 10×10 opto-electronic detector array at the bridge/junction between the two modalities (optical vs. electronic). (c) and (d) are same as (a) and (b), except the latter are for Fashion-MNIST dataset. Input plane represents the plane of the input object or its data, which can also be generated by another optical imaging system or a lens, projecting an image of the object data onto this plane.