Fig. 11From: Face image synthesis from facial partsApply the model trained on CelebA to the CACD dataset. The first row is randomly selected images from CACD, and the last row is a synthetic portrait of our model. (FP is short for facial parts.)Back to article page