MFOC-CliqueNet: A CliqueNet-Based Optimal Combination of Multidimensional Features Classification Method for Large-Scale Laser Point Clouds

Table 7

Methods comparison accuracy (%).

Class	Method
	Pole	Vegetation	Wire	Ground	Facade	OA
M3N [23]	28.7	97.4	12.5	98.2	90.8	91.66
Literature [24]	22.3	90.7	5.3	99.6	87.6	93.53
Literature [25]	70.11	80.55	93.08	98.22	70.95	94.68
FDM [18]	68.42	80.68	92.93	98.37	71.13	94.75
CRFoptN [26]	59.7	92.0	10.7	99.9	94.6	95.5
MRF [27]	68.0	95.5	51.3	98.4	92.9	97.0
CCM [28]	82.67	97.83	30.26	99.17	90.33	97.59
Our method	16.0	89.8	59.1	99.5	94.6	98.9

Bold values highlight the advantages of our method and other literature methods in classification accuracy comparison.