Sync4D: Video Guided Controllable Dynamics for Physics-Based 4D Generation

TL;DR

This paper presents Sync4D, a method for controllable, high-fidelity 4D dynamic generation of 3D Gaussians driven by reference videos, combining shape reconstruction, motion transfer, and physical simulation.

Contribution

It introduces a novel approach that transfers motion from reference videos to 3D Gaussians with shape and temporal consistency using shape correspondence and physical simulation.

Findings

Supports diverse reference inputs including humans and animals

Generates arbitrary-length dynamic sequences with high fidelity

Outperforms existing methods in motion accuracy and consistency

Abstract

In this work, we introduce a novel approach for creating controllable dynamics in 3D-generated Gaussians using casually captured reference videos. Our method transfers the motion of objects from reference videos to a variety of generated 3D Gaussians across different categories, ensuring precise and customizable motion transfer. We achieve this by employing blend skinning-based non-parametric shape reconstruction to extract the shape and motion of reference objects. This process involves segmenting the reference objects into motion-related parts based on skinning weights and establishing shape correspondences with generated target shapes. To address shape and temporal inconsistencies prevalent in existing methods, we integrate physical simulation, driving the target shapes with matched motion. This integration is optimized through a displacement loss to ensure reliable and genuine dynamics. Our approach supports diverse reference inputs, including humans, quadrupeds, and articulated objects, and can generate dynamics of arbitrary length, providing enhanced fidelity and applicability. Unlike methods heavily reliant on diffusion video generation models, our technique offers specific and high-quality motion transfer, maintaining both shape integrity and temporal consistency.