Nonlinear All-Optical Diffractive Deep Neural Network with 10.6 μm Wavelength for Image Classification

Table 10

The classification accuracy in the standard N-D²NN model corresponds to pixel size and diffraction grating spacing.

Accuracy (%)		Pixel size
		30 × 30	40 × 40	50 × 50	60 × 60	70 × 70	80 × 80	90 × 90	100 × 100
Spacing (λ)	30	84.27	86.42	87.36	86.94	86.64	86.56	86.52	86.50
	40	82.18	86.23	87.07	87.44	86.67	86.61	86.94	86.77
	50	75.45	86.14	85.94	87.59	87.12	86.72	86.83	87.03
	60	64.99	83.27	87.10	87.14	87.41	86.94	86.80	86.97
	70	61.90	83.04	86.83	87.65	86.96	86.88	86.75	86.59