Implementing PointNet for point cloud segmentation in the heritage context

Heritage Science

Table 2 Implementation results

Dataset	Building/ number of rooms	Dataset		Accuracy		Loss
Dataset	Building/ number of rooms	Train	Test	Train	Test	Train	Test
HBIM Gaziantep Cultural Heritage	18 HBIM building /100 room	Area_1 Area_2 Area_3 Area_4 Area_5	Area_6 (Laser Scanner data)	0.7147	0.5783	1.1179	1.8003
PointNet-Stanford	6 PointNet building / 271 room	Area_1 Area_2 Area_3 Area_4 Area_5	Area_6	0.8102	0.8200	0.7072	0.7526
HBIM Gaziantep Cultural Heritage	18 HBIM building / 100 room + 1 PointNet building / 40 room	Area_2 Area_3 Area_4 Area_5 Area_6	Area_1 (Laser Scanner data)	0.9230	0.8793	0.1818	0.4410
HBIM Gaziantep Cultural Heritage	18 HBIM building/ 100 room + 1 PointNet building /40 room + 11 restitution Building	Building_6 Building_7 … Building_21 Building_22	Building_30 (Restitution data)	0.9514	0.8652	0.3320	0.4150
			Building_29 (Restitution data)	0.9514	0.8159	0.3320	0.7734
			Building_25 (Restitution data)	0.9514	0.9120	0.3320	0.3250