Fig. 6

Comparison of diffractive network output images under different conditions. a Output images corresponding to the same input test object imaged through diffractive networks trained with n=1, n=10, n=15 and n=20. Second column: imaged through a known diffuser; third column: imaged through a new diffuser; fourth column: imaged without a diffuser. b The PCC values corresponding to the networks trained with n = 1, 10, 15 and 20 over input test objects distorted by known diffusers, new diffusers, as well as imaged without a diffuser. The error bars reflect the standard deviation over different diffusers