Human Pose Transfer by Adaptive Hierarchical Deformation
This work provides an incremental improvement for researchers and practitioners in computer vision focusing on realistic human image synthesis and animation.
This paper addresses the challenge of human pose transfer, which often struggles with preserving hair and clothing styles. The authors propose an adaptive hierarchical deformation network that generates a textured person image guided by semantic parsing, achieving better performance in consistency of hair, face, and clothes with fewer parameters than state-of-the-art methods.
Human pose transfer, as a misaligned image generation task, is very challenging. Existing methods cannot effectively utilize the input information, which often fail to preserve the style and shape of hair and clothes. In this paper, we propose an adaptive human pose transfer network with two hierarchical deformation levels. The first level generates human semantic parsing aligned with the target pose, and the second level generates the final textured person image in the target pose with the semantic guidance. To avoid the drawback of vanilla convolution that treats all the pixels as valid information, we use gated convolution in both two levels to dynamically select the important features and adaptively deform the image layer by layer. Our model has very few parameters and is fast to converge. Experimental results demonstrate that our model achieves better performance with more consistent hair, face and clothes with fewer parameters than state-of-the-art methods. Furthermore, our method can be applied to clothing texture transfer.