Frames with groundtruth hair mask annotation - Here're the groundtruth data we used in Fig 10.

Input video clips.

Output mesh results.