Skip to main content
. 2018 Nov 13;9(12):6205–6221. doi: 10.1364/BOE.9.006205

Fig. 5.

Fig. 5

Qualitative results for perceived clarity for 55 B-scans ranked by observer 1 (a), observer 2 (b) and observer 3 (c) for images acquired by 6-frame averaging (“Ground Truth”), CNN-WGAN, CNN-MSE, BM3D, DD-CDWT and the respective un-processed images (“Raw”). Colour indicates the percentage of images, for the respective image type, in each rank position (where 1=highest rank, 6=lowest rank). Results for accuracy and personal preference are shown in Figure 7.