April 2020
Intermediate to advanced
438 pages
12h 2m
English
Train a completion net using your own images (refer to the GitHub repository provided in the See also section for the code to train; you will need lots of images, a long time, and a GPU). Also, you can use the pre-trained Completion Network to complete (inpaint) faces too. Try it on your own—you should get an output as shown in the following screenshot:

Read now
Unlock full access