High-Level Interpretation of Urban Road Maps Fusing Deep Learning-Based Pixelwise Scene Segmentation and Digital Navigation Maps

Table 4

Quantitative results on F-score in KITTI dataset.

	Image plane				Bird Eye View
Classifier	UM	UMM	UU	All	UM	UMM	UU	All
Boosting	85.24	93.09	82.08	87.06	85.15	89.37	62.60	79.05
Map Model	89.35	88.56	90.20	89.37	85.17	84.76	86.66	85.53
CNN	96.64	95.36	94.89	95.50	95.09	94.94	93.39	94.59