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

Figure 9

Results of 3D detection on the KITTI test set: (a) the corresponding 2D image; (b) the detection results of the Pointpillar; (c) the detection results of the Pointpillar+SE; (d) the detection results of ours.

(a)

(b)

(c)

(d)