Automated summarisation of SDOCT volumes using deep learning: Transfer learning vs de novo trained networks - Fig 10

(a) The en-face projection image of the volume, with (b) all the key B-scans and (c) the final visual summary. The three B-scans corresponding (d)—(f) to the first region (blues), (g)—(i) the second region (reds), and (j)—(l) to the third region (greens).