Refining 3D Human Texture Estimation from a Single Image
This addresses the challenge of creating realistic 3D human models from limited 2D data for graphics and vision applications, representing an incremental advance with specific technical contributions.
The paper tackles the problem of estimating 3D human texture from a single image by proposing a framework with deformable convolution sampling, a cycle consistency loss, and an uncertainty-based reconstruction loss, showing significant qualitative and quantitative improvements over state-of-the-art methods.
Estimating 3D human texture from a single image is essential in graphics and vision. It requires learning a mapping function from input images of humans with diverse poses into the parametric (UV) space and reasonably hallucinating invisible parts. To achieve a high-quality 3D human texture estimation, we propose a framework that adaptively samples the input by a deformable convolution where offsets are learned via a deep neural network. Additionally, we describe a novel cycle consistency loss that improves view generalization. We further propose to train our framework with an uncertainty-based pixel-level image reconstruction loss, which enhances color fidelity. We compare our method against the state-of-the-art approaches and show significant qualitative and quantitative improvements.