Table 1 Quantitative comparison (MSE/PSNR/SSIM) of the proposed model with KL divergence loss
From: DEEP-squared: deep learning powered De-scattering with Excitation Patterning
Dataset | SLS | Evaluation metrics | ||
|---|---|---|---|---|
MSE (10−4) | PSNR | SSIM (102) | ||
Mouse cortical vasculature | 2 SLS | 4.33 ± 2.46 | 34.11 ± 1.87 | 92.45 ± 1.97 |
4 SLS | 6.77 ± 3.01 | 32.03 ± 1.62 | 88.79 ± 2.08 | |
6 SLS | 29.07 ± 12.71 | 25.72 ± 1.72 | 80.76 ± 3.86 | |
Mouse pyramidal neuron | 2 SLS | 2.76 ± 1.49 | 36.18 ± 2.28 | 97.10 ± 1.29 |
6 SLS | 15.46 ± 9.51 | 28.99 ± 2.89 | 85.81 ± 4.74 | |
Synthetic beads | 4 SLS | 1.33 ± 0.58 | 39.16 ± 1.82 | 78.75 ± 12.95 |
