Image2Mesh: A Learning Framework for Single Image 3D Reconstruction

TL;DR

This paper introduces a novel learning framework that reconstructs detailed 3D meshes from single images using a compact, low-dimensional representation, avoiding reliance on silhouettes or landmarks.

Contribution

It proposes a new mesh inference method based on free-form deformation and sparse linear combinations, improving 3D reconstruction accuracy and efficiency.

Findings

Effective reconstruction of 3D meshes from single images

Preserves fine geometric details in reconstructions

Works well on both synthetic and real-world data

Abstract

One challenge that remains open in 3D deep learning is how to efficiently represent 3D data to feed deep networks. Recent works have relied on volumetric or point cloud representations, but such approaches suffer from a number of issues such as computational complexity, unordered data, and lack of finer geometry. This paper demonstrates that a mesh representation (i.e. vertices and faces to form polygonal surfaces) is able to capture fine-grained geometry for 3D reconstruction tasks. A mesh however is also unstructured data similar to point clouds. We address this problem by proposing a learning framework to infer the parameters of a compact mesh representation rather than learning from the mesh itself. This compact representation encodes a mesh using free-form deformation and a sparse linear combination of models allowing us to reconstruct 3D meshes from single images. In contrast to prior work, we do not rely on silhouettes and landmarks to perform 3D reconstruction. We evaluate our method on synthetic and real-world datasets with very promising results. Our framework efficiently reconstructs 3D objects in a low-dimensional way while preserving its important geometrical aspects.