Multiscale Convolutional Neural Networks for Hand Detection

Table 4

Average Recall (AR) on VIVA L1 and L2 dataset and comparison with previous methods.

Method	L1 set	L2 set
CNNRegionSampling [9]	48.1%	36.6%
ACF Depth4 [6]	53.8%	40.4%
YOLO [10]	46.0%	39.1%
FRCNN [11]	55.9%	53.3%
Our model (Multiscale Fast R-CNN)	82.8%	66.5%