Pillar-Based 3D Object Detection from Point Cloud with Multiattention Mechanism

Table 9

Results on the KITTI test average orientation similarity (AOS) detection benchmark.

Model	Speed (Hz)	mAOS	Car			Pedestrian			Cyclist
		Mod.	Easy (%)	Mod. (%)	Hard (%)	Easy (%)	Mod. (%)	Hard (%)	Easy (%)	Mod. (%)	Hard (%)
SubCNN [42]	0.5	72.71	90.61	88.43	78.63	78.33	66.28	61.37	71.39	63.41	56.34
AVOD-FPN [39]	10	63.19	89.95	87.13	79.74	53.36	44.92	43.77	67.61	57.53	54.16
SECOND	20	54.53	87.84	81.31	71.95	51.56	43.51	38.78	80.97	57.20	55.14
PointPillars	62	68.86	90.19	88.76	86.38	32.39	31.41	29.84	82.43	68.16	61.96
Ours	34	72.14	94.89	91.09	88.47	56.52	50.99	47.57	86.80	74.34	70.67