Skip to main content

Table 3 Reported training and testing times for various algorithms on MSRC-21.

From: Scene Segmentation with Low-Dimensional Semantic Representations and Conditional Random Fields

Method

Training time

Test time

TextonBoost [7]

2 days

30 sec/image

PLSA-MRF [13]

1 hour

2 sec/image

STF-ILP [19]

2 hours

<0.125 sec/image

AC(ACP) [18]

a few days

30–70 sec/image

Our LRC/LRC

7-8 min

<0.03 sec/image

Our LRC/CRF

30–35 min

<0.02 sec/image