Research Article

Deep Learning for Person Reidentification Using Support Vector Machines

Table 2

Comparison of state-of-the-art results of feature representation reported with VIPeR database. The cumulative matching scores (%) at ranks 1, 5, 10, and 20 are listed.

MethodVIPeR
Top 1Top 5Top 10Top 20

ELF68.7318.7623.7531.75
gBiCov9.8727.6436.7548.96
HSV_Lab_LBP12.4726.9533.3744.16
Ours34.1567.8680.9590.63