MANet: End-to-End Learning for Point Cloud Based on Robust Pointpillar and Multiattention

Table 1

Ablation experiments on the effect of CBAM, SE, and TA as well as different combination settings.

Methods	Car	Pedestrian	Cyclist	mAP
Baseline [6]	74.9	43.5	64.5	59.0
SE [18]	68.3	44.94	51.72	54.9
CBAM [8]	76.1	50.2	64.3	63.5
TA [7]	78.8	46.47	59.60	62
TA + CBAM	78.8	51.74	64.74	65.1