Incremental Instance-Oriented 3D Semantic Mapping via RGB-D Cameras for Unknown Indoor Scene

Table 3

Comparison to the 3D semantic instance-segmentation approach from Voxblox++ [16] and Pham et al. [45] on class-averaged mAP value.

Sequence ID	Voxblox++ [16]	Pham et al. [45]	Ours
011	75.0	52.1	78.6
016	33.3	34.2	25.0
030	56.1	56.8	58.6
061	66.7	59.1	46.6
078	45.2	34.9	69.8
086	20.0	35.0	47.2
096	29.2	26.5	26.7
206	79.6	41.7	78.0
223	43.8	40.9	45.8
255	75.0	48.6	75.0