Global motion estimation (GME) plays an important role in video object segmentation. This paper presents a computationally efficient three stage affine GME algorithm, using radius-based Fourier Descriptors from histogram-based image segmented regions. Then a variance-cut KD tree is used for initial matching between FDs. and an efficient two-step outlier method is applied to remove incorrect outliers. Experiments with different video sequences are used to demonstrate the performance of the proposed approach.
展开▼
