STCT: Sequentially Training Convolutional Networks for Visual Tracking | IEEE Conference Publication | IEEE Xplore