Vehicle Reidentification Based on MAPANet and k-Reciprocal Encoding

Table 3

The Rank-1 and Rank-5 comparison on the VehicleID dataset.

Method	Small (800)		Medium (1600)		Large (2400)
	R1 (%)	R5 (%)	R1 (%)	R5 (%)	R1 (%)	R5 (%)
GoogLeNet [32]	47.88	67.18	43.40	63.86	38.27	59.39
FACT [10]	49.53	68.07	44.59	64.57	39.92	60.32
XVGAN [41]	52.79	80.69	49.47	71.42	44.92	66.71
VRSDNet [37]	56.98	86.90	50.57	80.05	42.92	73.44
VAMI [13]	63.12	83.25	52.87	75.12	47.34	70.29
DLCNN [42]	73.01	82.70	66.50	77.06	61.00	73.17
EALN [40]	75.11	—	78.69	—	69.30	—
GS-TRE [43]	75.90	84.20	74.80	83.60	72.30	82.70
DFN [26]	77.02	85.04	71.81	80.81	66.29	78.42
Ours	78.15	89.92	74.21	84.86	72.31	84.15