RGBD Scene Flow Estimation with Global Nonrigid and Local Rigid Assumption

<div>Pose 1 in “Hat” sequence.” Input two consecutive frames about Pose 1 in “Hat” sequence, segment them into 2 regions (human body with hat and background), and estimate the occlusion and scene flow. (a) RGB images. (b) Depth images. (c) Segmentation <i>K</i> = 2. (d) Occlusions. (e) Motion.</div>

Discrete Dynamics in Nature and Society

fig6

Figure 6

Figure 6: RGBD Scene Flow Estimation with Global Nonrigid and Local Rigid Assumption 