Self-Calibrating 4D Novel View Synthesis from Monocular Videos Using Gaussian Splatting

TL;DR

This paper introduces a self-calibrating 4D Gaussian Splatting method for monocular videos that improves scene reconstruction and view synthesis accuracy without relying on external camera parameter estimates.

Contribution

It presents a novel approach for joint optimization of camera parameters and scene structure using 2D point features, enhancing robustness and efficiency in 4D scene reconstruction.

Findings

Significant accuracy improvements over state-of-the-art methods.

Enhanced robustness in scenes with large object movements.

Faster reconstruction times demonstrated on standard benchmarks.

Abstract

Gaussian Splatting (GS) has significantly elevated scene reconstruction efficiency and novel view synthesis (NVS) accuracy compared to Neural Radiance Fields (NeRF), particularly for dynamic scenes. However, current 4D NVS methods, whether based on GS or NeRF, primarily rely on camera parameters provided by COLMAP and even utilize sparse point clouds generated by COLMAP for initialization, which lack accuracy as well are time-consuming. This sometimes results in poor dynamic scene representation, especially in scenes with large object movements, or extreme camera conditions e.g. small translations combined with large rotations. Some studies simultaneously optimize the estimation of camera parameters and scenes, supervised by additional information like depth, optical flow, etc. obtained from off-the-shelf models. Using this unverified information as ground truth can reduce robustness and accuracy, which does frequently occur for long monocular videos (with e.g. > hundreds of frames). We propose a novel approach that learns a high-fidelity 4D GS scene representation with self-calibration of camera parameters. It includes the extraction of 2D point features that robustly represent 3D structure, and their use for subsequent joint optimization of camera parameters and 3D structure towards overall 4D scene optimization. We demonstrate the accuracy and time efficiency of our method through extensive quantitative and qualitative experimental results on several standard benchmarks. The results show significant improvements over state-of-the-art methods for 4D novel view synthesis. The source code will be released soon at https://github.com/fangli333/SC-4DGS.