Skip to main content
. 2018 Apr 17;18(4):1232. doi: 10.3390/s18041232

Figure 8.

Figure 8

Comparison of reconstruction performances. The reconstructed image from ICNET (d) has a larger number of differences in the foreground region than AE (b). These larger differences lead to a larger recall in ICNET-ED, as shown in Figure 7.