. Author manuscript; available in PMC: 2023 Apr 1.

Published in final edited form as: IEEE Trans Pattern Anal Mach Intell. 2022 Mar 4;44(4):1934–1948. doi: 10.1109/TPAMI.2020.3033882

TABLE 5. Average Metrics on cropped Display Captured PhlatCam measurements.

FlatNet-gen performs consistently better than other learned approaches for both real (FlatNet-gen-C) and simulated PSF case(FlatNet-gen-UC). It should be noted that FlatNet-gen-UC performs as good as Le-ADMM based on real PSF.

Method	PSNR(in dB)	SSIM	LPIPS
Tikh+U-Net-UC	17.53	0.45	0.438
Tikh+U-Net-C	18.34	0.48	0.376
Le-ADMM-UC	17.94	0.45	0.410
Le-ADMM-C	18.72	0.48	0.371
FlatNet-gen-UC	18.72	0.48	0.375
FlatNet-gen-C	19.29	0.50	0.365