4DPV: 4D Pet from Videos by Coarse-to-Fine Non-Rigid Radiance Fields

TL;DR

This paper introduces a self-supervised, coarse-to-fine neural deformation framework for 4D reconstruction and pose estimation of unknown objects from videos, without requiring pre-existing 3D data or templates.

Contribution

It proposes a novel neural deformation model that captures both coarse and fine details using canonical and image-variant spaces, enabling accurate 4D reconstruction from in-the-wild videos.

Findings

Effective on complex real-world deformations

Outperforms existing approaches in accuracy

Provides detailed qualitative and quantitative results

Abstract

We present a coarse-to-fine neural deformation model to simultaneously recover the camera pose and the 4D reconstruction of an unknown object from multiple RGB sequences in the wild. To that end, our approach does not consider any pre-built 3D template nor 3D training data as well as controlled illumination conditions, and can sort out the problem in a self-supervised manner. Our model exploits canonical and image-variant spaces where both coarse and fine components are considered. We introduce a neural local quadratic model with spatio-temporal consistency to encode fine details that is combined with canonical embeddings in order to establish correspondences across sequences. We thoroughly validate the method on challenging scenarios with complex and real-world deformations, providing both quantitative and qualitative evaluations, an ablation study and a comparison with respect to competing approaches. Our project is available at https://github.com/smontode24/4DPV.