Skip to main content
. 2021 Jun 16;12(6):702. doi: 10.3390/mi12060702

Figure 17.

Figure 17

Train result (under the condition of trial 1): (a) loss of “sequence to sequence” model; (b) loss of “image to sequence” model.