PPANet: Point-Wise Pyramid Attention Network for Semantic Segmentation

Table 3

Comparison of our model with other networks on the KITTI road segmentation dataset, using average precision (AP), precision, recall, and -measure.

Method	AP (%)	Pre (%)	Rec (%)
SegNet [23]	89.40	90.90	89.50	90.10
ENet [21]	87.40	88.90	87.60	88.20
FastFCN [68]	95.10	96.10	94.70	95.30
LBN-AA [66]	94.70	96.00	94.10	95.00
DABNet [67]	93.40	94.90	93.50	94.10
AGLNet [69]	94.60	96.00	95.60	96.30
Ours (PPANet)	96.20	97.10	95.60	96.30