Skip to main content
. 2020 Jan 15;10:338. doi: 10.1038/s41598-019-56989-5

Table 7.

Visual assessment of the proposed ensemble and local histogram-based method. ECNN stands for the proposed Ensemble of Convolutional Networks method.

MPV MAE Over. Under.
(mean ± sd) LH ECNN LH ECNN LH ECNN
NP 73.2 ± 10.9 1.25 0.80 6.00 5.60 3.89 4.37
RETIC 10.0 ± 5.6 2.45 0.80 7.45 5.77 5.00 4.75
PS 3.2 ± 1.6 0.70 0.70 5.07 5.50 4.32 4.46
CLE 6.5 ± 9.2 1.45 0.80 6.33 5.33 4.17 4.23
SUBPL 1.5 ± 0.6 0.75 0.25 5.75 5.21 5.00 4.94
GG 1.2 ± 1.0 0.40 0.50 5.40 5.07 5.00 4.52
NOD 2.5 ± 3.8 0.00 1.15 5.00 6.15 5.00 5.00
LINSC 1.9 ± 0.9 1.40 0.70 6.37 5.70 4.78 5.00
Sp. Cons.* 2.6 1.10
H. Prec.** 2.35 0.85
Mean 1.33 0.76 5.92 5.54 4.64 4.66

*Spatial Consistency; **Hilar Precision.

MPV: Mean Percentage Volume, MAE: Mean Absolute Error, Over.: Overestimation, Under.: Underestimation.

LH stands for Local Histogram.