Fig. 3: Generalization of FIN on different types of training samples. | Light: Science & Applications

Fig. 3: Generalization of FIN on different types of training samples.

From: Fourier Imager Network (FIN): A deep neural network for hologram reconstruction with superior external generalization

Fig. 3

a Each FIN model is separately trained using the raw holograms (M = 3) of the corresponding sample type. Scale bar: 30 μm. b A comparison of the amplitude and phase SSIM values corresponding to the hologram reconstruction performed by different FIN models that are trained and tested with different types of samples. The ground truth for each sample is obtained through the MH-PR algorithm that used M = 8 raw holograms captured at different sample-to-sensor distances

Back to article page