This thesis focuses on the advancement of the Discriminative Correlation Filter (DCF) framework for visual tracking, which is one of the fundamental problems in computer vision. The main contribution of this thesis is the study of efficient update rules and numerical solvers for the learning of the appearance model, which is used to estimate the target location in a sequence of images. Additionally, the periodic assumption induced by the circular convolution in DCF is countered by proposing a novel approach.
Previous ArticleMachine Learning Revolutionizes Ecg Diagnosis For Heart Attacks
Next Article New Technology Jobs Needed For Ai