Delving StyleGAN Inversion for Image Editing: A Foundation Latent Space Viewpoint

Figure 1 . The inversion and editing results of our model in the real images. We show from the left to right of each row: an input image, inversion results, and our editing results. We edit images by modifying the attributes in the embedding space following [22, 56] . The ↓ means a decreased magnitude of the manipulation attribute.