Large-Scale Video Retrieval via Deep Local Convolutional Features

Table 6

Comparison with existing visual search approaches on the 2001 TREC dataset.

Method	Acc	Time (s)	Dimension
Ours	0.9213	5.302	64, 512
DFSP	0.8841	5.164	64
DF [16]	0.8313	14.305	25088
DFSE [27]	0.8106	5.201	64
DFTA [28]	0.8027	8.132	512
LBTT [28]	0.8174	6.764	256