Inferring and Leveraging Parts from Object Shape for Improving Semantic Image Synthesis

Ours support PartNet (d) Part Editing Figure 1. Illustration of our iPOSE for part map prediction and semantic image synthesis. (a) Calculation of SPD [25]. (b) Our PartNet for part map prediction. (c) Comparison of results with existing methods [25, 34]. (d) Part editing results of our iPOSE.