Figure 4
From: Deep learning alignment of bidirectional raster scanning in high speed photoacoustic microscopy
Extrapolation application of MS-FD-U-Net. (a–c) Depth-encoded PA images of a mouse’s ear reconstructed with (a) bidirectionally (marked as Input) and (b) unidirectionally (marked as Ground truth) acquired data, (c) the MS-FD-U-Net GAN. Close-up PA MAP views of ROI 1 and 2 in the regions indicated by the blue and green boxes are presented, respectively. PA amplitude profiles indicated by the dotted lines in (d) ROI 1 and (e) ROI 2, respectively. MAP, maximum amplitude projection; ROI, region of interest; and Norm. PA amp., normalized PA amplitude.
