Synthesizing Anyone, Anywhere, in Any Pose
This addresses the problem of synthesizing realistic human figures in complex, uncontrolled environments for applications like image editing and virtual reality, representing a strong specific gain in a domain-specific area.
The paper tackles the challenging task of in-the-wild human figure synthesis, aiming to generate full-body images from any region in any image, and demonstrates that TriA-GAN significantly improves over previous methods while requiring less conditional information.
We address the task of in-the-wild human figure synthesis, where the primary goal is to synthesize a full body given any region in any image. In-the-wild human figure synthesis has long been a challenging and under-explored task, where current methods struggle to handle extreme poses, occluding objects, and complex backgrounds. Our main contribution is TriA-GAN, a keypoint-guided GAN that can synthesize Anyone, Anywhere, in Any given pose. Key to our method is projected GANs combined with a well-crafted training strategy, where our simple generator architecture can successfully handle the challenges of in-the-wild full-body synthesis. We show that TriA-GAN significantly improves over previous in-the-wild full-body synthesis methods, all while requiring less conditional information for synthesis (keypoints \vs DensePose). Finally, we show that the latent space of TriA-GAN is compatible with standard unconditional editing techniques, enabling text-guided editing of generated human figures.