Learned Spatial Representations for Few-shot Talking-Head Synthesis
This work addresses identity preservation in talking-head synthesis, which is crucial for applications like video conferencing and entertainment, though it appears incremental as it builds on existing neural talking-head approaches.
The paper tackles the problem of preserving subject identity in few-shot talking-head synthesis by factorizing representations into spatial and style components, leading to significant quantitative and qualitative improvements over previous methods.
We propose a novel approach for few-shot talking-head synthesis. While recent works in neural talking heads have produced promising results, they can still produce images that do not preserve the identity of the subject in source images. We posit this is a result of the entangled representation of each subject in a single latent code that models 3D shape information, identity cues, colors, lighting and even background details. In contrast, we propose to factorize the representation of a subject into its spatial and style components. Our method generates a target frame in two steps. First, it predicts a dense spatial layout for the target image. Second, an image generator utilizes the predicted layout for spatial denormalization and synthesizes the target frame. We experimentally show that this disentangled representation leads to a significant improvement over previous methods, both quantitatively and qualitatively.