Fig. 3

Exploration of the learned mapping of the generator model. A Examples of simulated (in silico), translated tissue velocity images, B their corresponding differences at three different time frames of an image sequence, and C is the similarity map computed as the sum of difference maps of the whole image sequence. D Similarity maps for three translated examples. E Similarity as a function of depth for all 64 translated examples. The similarity was low in superficial subcutaneous region. F Examples of experimental tissue velocity frames. It can be seen by visual inspection that the texture pattern of the similarity maps (D) shared similar features with the texture pattern of the experimental velocity maps (F)