Closest to our goal of simultaneously ob- taining segmentations and 3D poses is Posecut[4], a pose estimation method which alternates between fitting a 3D.
確定! 回上一頁