Skip to main content
. 2018 Dec 24;19(1):59. doi: 10.3390/s19010059

Table 9.

Test results of three investigated CNN models on a total of four different datasets in terms of classification error.

W150 W100 + 50 ROTARM REHAP
CNN-2DP 36.1% 41.8% 62.0% 16.1%
CNN-PPC 17.5% 27.4% 55.6% 16.5%
CNN-NPC 39.5% 40.2% 73.3% 25.2%