NGD: Neural Gradient Based Deformation for Monocular Garment Reconstruction
This addresses the problem of reconstructing dynamically evolving garments from monocular video for computer vision and graphics applications, representing a strong domain-specific advancement.
The paper tackles dynamic garment reconstruction from monocular video by proposing NGD, a neural gradient-based deformation method that addresses limitations in existing implicit and template-based approaches. The method achieves significant improvements over state-of-the-art methods, providing high-quality reconstructions with detailed wrinkles and pleats.
Dynamic garment reconstruction from monocular video is an important yet challenging task due to the complex dynamics and unconstrained nature of the garments. Recent advancements in neural rendering have enabled high-quality geometric reconstruction with image/video supervision. However, implicit representation methods that use volume rendering often provide smooth geometry and fail to model high-frequency details. While template reconstruction methods model explicit geometry, they use vertex displacement for deformation, which results in artifacts. Addressing these limitations, we propose NGD, a Neural Gradient-based Deformation method to reconstruct dynamically evolving textured garments from monocular videos. Additionally, we propose a novel adaptive remeshing strategy for modelling dynamically evolving surfaces like wrinkles and pleats of the skirt, leading to high-quality reconstruction. Finally, we learn dynamic texture maps to capture per-frame lighting and shadow effects. We provide extensive qualitative and quantitative evaluations to demonstrate significant improvements over existing SOTA methods and provide high-quality garment reconstructions.