Efficiently Reconstructing Dynamic Scenes One D4RT at a Time

TL;DR

This paper presents D4RT, a transformer-based model that efficiently reconstructs dynamic scenes from video by jointly estimating depth, motion, and camera parameters, achieving state-of-the-art results with a scalable and lightweight approach.

Contribution

Introducing D4RT, a unified transformer architecture with a novel querying mechanism for efficient 4D scene reconstruction from video.

Findings

Outperforms previous methods across multiple 4D reconstruction tasks.

Enables efficient training and inference due to its lightweight design.

Achieves state-of-the-art accuracy in dynamic scene reconstruction.

Abstract

Understanding and reconstructing the complex geometry and motion of dynamic scenes from video remains a formidable challenge in computer vision. This paper introduces D4RT, a simple yet powerful feedforward model designed to efficiently solve this task. D4RT utilizes a unified transformer architecture to jointly infer depth, spatio-temporal correspondence, and full camera parameters from a single video. Its core innovation is a novel querying mechanism that sidesteps the heavy computation of dense, per-frame decoding and the complexity of managing multiple, task-specific decoders. Our decoding interface allows the model to independently and flexibly probe the 3D position of any point in space and time. The result is a lightweight and highly scalable method that enables remarkably efficient training and inference. We demonstrate that our approach sets a new state of the art, outperforming previous methods across a wide spectrum of 4D reconstruction tasks. We refer to the project webpage for animated results: https://d4rt-paper.github.io/.