We randomly selected 30 samples from the InterHuman dataset for the user study. Shown here are the generated results from HINT (A), DART$^{\dagger}$ (B), and InterMask (C).