Accelerate Learning of Deep Hashing With Gradient Attention | IEEE Conference Publication | IEEE Xplore