CLIP-Guided StyleGAN Inversion for Text-Driven Real Image Editing
This work addresses a specific bottleneck in text-driven image editing for applications like media and design, offering an incremental improvement over prior StyleGAN-based approaches.
The paper tackles the problem of inefficient and ineffective multi-attribute text-driven image editing by proposing CLIPInverter, which uses text-conditioned adapter layers and CLIP-guided refinement to achieve more accurate and photo-realistic edits, outperforming existing methods in manipulation accuracy and photo-realism across domains like human faces, cats, and birds.
Researchers have recently begun exploring the use of StyleGAN-based models for real image editing. One particularly interesting application is using natural language descriptions to guide the editing process. Existing approaches for editing images using language either resort to instance-level latent code optimization or map predefined text prompts to some editing directions in the latent space. However, these approaches have inherent limitations. The former is not very efficient, while the latter often struggles to effectively handle multi-attribute changes. To address these weaknesses, we present CLIPInverter, a new text-driven image editing approach that is able to efficiently and reliably perform multi-attribute changes. The core of our method is the use of novel, lightweight text-conditioned adapter layers integrated into pretrained GAN-inversion networks. We demonstrate that by conditioning the initial inversion step on the CLIP embedding of the target description, we are able to obtain more successful edit directions. Additionally, we use a CLIP-guided refinement step to make corrections in the resulting residual latent codes, which further improves the alignment with the text prompt. Our method outperforms competing approaches in terms of manipulation accuracy and photo-realism on various domains including human faces, cats, and birds, as shown by our qualitative and quantitative results.