StyleGANEX: StyleGAN-Based Manipulation Beyond Cropped Aligned Faces
It addresses the problem of limited face manipulation flexibility for researchers and practitioners by enabling StyleGAN-based methods to handle unaligned faces, though it is incremental as it builds on existing StyleGAN architecture.
The paper tackles StyleGAN's limitation to cropped aligned faces by using dilated convolutions to extend shallow-layer receptive fields, enabling manipulation of unaligned faces at variable resolutions without altering model parameters, and validates this with tasks like attribute editing and super-resolution.
Recent advances in face manipulation using StyleGAN have produced impressive results. However, StyleGAN is inherently limited to cropped aligned faces at a fixed image resolution it is pre-trained on. In this paper, we propose a simple and effective solution to this limitation by using dilated convolutions to rescale the receptive fields of shallow layers in StyleGAN, without altering any model parameters. This allows fixed-size small features at shallow layers to be extended into larger ones that can accommodate variable resolutions, making them more robust in characterizing unaligned faces. To enable real face inversion and manipulation, we introduce a corresponding encoder that provides the first-layer feature of the extended StyleGAN in addition to the latent style code. We validate the effectiveness of our method using unaligned face inputs of various resolutions in a diverse set of face manipulation tasks, including facial attribute editing, super-resolution, sketch/mask-to-face translation, and face toonification.