Skip to main content

Table 4 Comparison of the time consumption of different models in the CPU training environment

From: Ancient mural segmentation based on a deep separable convolution network

Model

Predict time

SegNet

33.4 s

PSPNet

28.4 s

DeeplabV3 + 

34.9 s

MC-DM

29.4 s

FCM

35.7 s

Grapth Cuts

2.3 s

PSP-M

17.4 s