Vehicle Attribute Recognition for Normal Targets and Small Targets Based on Multitask Cascaded Network

Table 5

The attribute recognition results for the normal targets on the SYIT-Vehicle dataset.

Method	Bus (%)	Microbus (%)	Minivan (%)	Sedan (%)	SUV (%)	Truck (%)	Total (%)
CNN	80.00	68.65	75.14	63.25	64.33	68.65	70.00
Fast R-CNN	78.92	70.81	73.51	68.33	70.27	76.22	72.97
Faster R-CNN	87.03	76.76	78.92	69.73	81.62	77.30	79.82
CNN + CNN	88.94	79.31	80.14	75.23	83.20	79.41	81.89
IFR-CNN + CNN	92.49	84.23	83.74	82.23	84.02	81.41	84.96
MC-CNN	91.35	83.70	83.70	80.92	84.78	81.00	84.24
MC-CNN-P	94.05	84.67	83.62	83.16	84.78	80.46	85.12
MC-CNN-F	94.60	84.62	84.24	84.08	87.57	83.62	86.46
MC-CNN-NT	97.30	91.11	90.57	88.41	90.73	89.49	91.27