Multiperspective Light Field Reconstruction Method via Transfer Reinforcement Learning

Table 3

Detection rate evaluation report form for different algorithms in the human dataset.

Method	Male, average	Female, average	Baby, average	Human dataset, average	Standard deviation
RPN	71.8	72.3	71.4	71.8	0.45
Fast R-CNN	71.3	71.3	71.1	71.2	0.12
YOLO 9000	70	71.1	69.9	70.3	0.67
SSD	70	69.8	70.3	70	0.25
R-CNN	69.8	70.7	70.1	70.2	0.46
Ours	73.6	72.5	71.4	72.5	1.1

VOC2012 human dataset detected accuracy (%). Average: average recognition rate of five samples in each dataset.