Fig. 9From: Learning attention for object tracking with adversarial learning networkThe maps generated by the proposed spatial attention scheme (middle row) and saliency detection algorithm [63, 64] Deep Saliency (bottom row)Back to article page