Our Semantic-based Cinemagraph Results

Input video

Ours (person)

Ours (tree)

Ours (grass)