Expressive Talking Head Video Encoding in StyleGAN2 Latent-Space
This work addresses the challenge of generating realistic animated face videos with detailed expressions for applications like video reenactment and puppeteering, representing an incremental improvement over existing methods.
The paper tackles the problem of capturing fine, expressive facial features in video reenactment by proposing an end-to-end encoding approach in StyleGAN2's latent space, achieving high-fidelity video rendering at 1024^2 resolution with data efficiency using a single identity latent and 35 parameters per frame.
While the recent advances in research on video reenactment have yielded promising results, the approaches fall short in capturing the fine, detailed, and expressive facial features (e.g., lip-pressing, mouth puckering, mouth gaping, and wrinkles) which are crucial in generating realistic animated face videos. To this end, we propose an end-to-end expressive face video encoding approach that facilitates data-efficient high-quality video re-synthesis by optimizing low-dimensional edits of a single Identity-latent. The approach builds on StyleGAN2 image inversion and multi-stage non-linear latent-space editing to generate videos that are nearly comparable to input videos. While existing StyleGAN latent-based editing techniques focus on simply generating plausible edits of static images, we automate the latent-space editing to capture the fine expressive facial deformations in a sequence of frames using an encoding that resides in the Style-latent-space (StyleSpace) of StyleGAN2. The encoding thus obtained could be super-imposed on a single Identity-latent to facilitate re-enactment of face videos at $1024^2$. The proposed framework economically captures face identity, head-pose, and complex expressive facial motions at fine levels, and thereby bypasses training, person modeling, dependence on landmarks/ keypoints, and low-resolution synthesis which tend to hamper most re-enactment approaches. The approach is designed with maximum data efficiency, where a single $W+$ latent and 35 parameters per frame enable high-fidelity video rendering. This pipeline can also be used for puppeteering (i.e., motion transfer).