Table 8. Comparative experimental results adopting ResNet101 as the backbone on the CULane dataset.

Bold numbers are the best results.

Method	Backbone	F1 (%)	FPS	Normal (%)	Crowd (%)	Dazzle (%)	Shadow (%)	No line (%)	Arrow (%)	Curve (%)	Cross	Night (%)
SCNN (Pan et al., 2018)	VGG16	71.60	7.5	90.60	69.70	58.50	66.90	43.40	84.10	64.40	1990	66.10
LaneAF (Abualsaud et al., 2021)	ERFNet	75.63	24	91.10	73.32	69.71	75.81	50.62	86.86	65.02	1844	70.90
LaneAF (Abualsaud et al., 2021)	DLA-34	77.41	20	91.80	75.61	71.78	79.12	51.38	86.88	72.70	1360	73.03
FOLOLane (Qu et al., 2021)	ERFNet	78.80	40	92.70	77.80	75.20	79.30	52.10	89.00	69.40	1569	74.50
LaneATT (Tabelini et al., 2021a)	Resnet122	77.02	26	91.74	76.16	69.47	76.31	50.46	86.29	64.05	1264	70.81
GANet-L (Morley et al., 2001)	Resnet101	79.63	63	93.67	78.66	71.82	78.32	53.38	89.86	77.37	1352	73.85
CondLane (Liu et al., 2021a)	Resnet101	79.48	47	93.47	77.44	70.93	80.91	54.13	90.16	75.21	1201	74.80
CLRNet (Zheng et al., 2022)	Resnet101	80.13	46	93.85	78.78	72.49	82.33	54.50	89.79	75.57	1262	75.51
Our method	Resnet101	80.31	67	94.04	78.91	74.64	82.56	54.69	89.84	75.83	1179	75.43