Abstract:This paper proposes a technique for efficiently modeling dynamic humans by explicifying the implicit neural fields via a Neural Explicit Surface (NES). Implicit neural fields have advantages over traditional explicit representations in modeling dynamic 3D content from sparse observations and effectively representing complex geometries and appearances. Implicit neural fields defined in 3D space, however, are expensive to render due to the need for dense sampling during volumetric rendering. Moreover, their memory efficiency can be further optimized when modeling sparse 3D space. To overcome these issues, the paper proposes utilizing Neural Explicit Surface (NES) to explicitly represent implicit neural fields, facilitating memory and computational efficiency. To achieve this, the paper creates a fully differentiable conversion between the implicit neural fields and the explicit rendering interface of NES, leveraging the strengths of both implicit and explicit approaches. This conversion enables effective training of the hybrid representation using implicit methods and efficient rendering by integrating the explicit rendering interface with a newly proposed rasterization-based neural renderer that only incurs a texture color query once for the initial ray interaction with the explicit surface, resulting in improved inference efficiency. NES describes dynamic human geometries with pose-dependent neural implicit surface deformation fields and their dynamic neural textures both in 2D space, which is a more memory-efficient alternative to traditional 3D methods, reducing redundancy and computational load. The comprehensive experiments show that NES performs similarly to previous 3D approaches, with greatly improved rendering speed and reduced memory cost.

Layered Depth-Normal Images: a Sparse Implicit Representation of Solid Models

Sketch-based 3-D modeling for piecewise planar objects in single images

A Hardware Assisted Ldi Building Algorithm With Application To Hybrid Volume Rendering

Sketch-based 3D Modeling by Aligning Outlines of an Image

Generation of Truss-Structure Objects with Implicit Representation for 3D-Printing

A New Framework for Layered Depth Video Coding

Parallel and efficient Boolean on polygonal solids

Layered Infill Area Generation from Triply Periodic Minimal Surfaces for Additive Manufacturing

Anti-Aliased Neural Implicit Surfaces with Encoding Level of Detail

Latent Partition Implicit with Surface Codes for 3D Representation

PolyDepth

Neighborhood Aided Implicit Active Contours

Layered Scene Models from Single Hazy Images.

Local Deep Implicit Functions for 3D Shape

Generation of Layered Depth Images from Multi-View Video

NESI: Shape Representation via Neural Explicit Surface Intersection

A Flexible Hierarchical Framework for Implicit 3D Characterization of Bionic Devices

A Latent Implicit 3D Shape Model for Multiple Levels of Detail

Explicifying Neural Implicit Fields for Efficient Dynamic Human Avatar Modeling via a Neural Explicit Surface

Multiresolution Normal Mesh Approximation of Implicit Surfaces

Enhancements of Representation and Interactivity for Multi-view Video Based on Layered Depth Image