MoRel: Long-Range Flicker-Free 4D Motion Modeling via Anchor Relay-based Bidirectional Blending with Hierarchical Densification
Sangwoon Kwak, Weeyoung Kwon, Jun Young Jeong, Geonho Kim, Won-Sik Cheong, Jihyong Oh
2025-12-12
Summary
This paper introduces a new technique, called MoRel, for creating realistic and fast-rendering videos of dynamic scenes, building on recent advances in representing 3D scenes with Gaussian splats.
What's the problem?
Existing methods for making these dynamic videos struggle when dealing with scenes that have a lot of movement over a long period of time. They require a huge amount of computer memory, often show distracting flickering, and have trouble with objects appearing or disappearing during the video.
What's the solution?
MoRel solves this by creating a system of 'anchor frames' at key moments in the video. It then tracks how things change between these anchors, instead of trying to track everything frame-by-frame. This reduces memory usage and creates smoother transitions. The system also intelligently adds more detail where it's needed, based on how much things are changing, and blends frames together in a way that minimizes flickering.
Why it matters?
This work is important because it allows for the creation of much more complex and realistic dynamic videos using Gaussian splats, without running into the limitations of previous methods. The researchers even created a new dataset with more challenging long-range motion to test their system, showing it can handle real-world scenarios effectively and efficiently.
Abstract
Recent advances in 4D Gaussian Splatting (4DGS) have extended the high-speed rendering capability of 3D Gaussian Splatting (3DGS) into the temporal domain, enabling real-time rendering of dynamic scenes. However, one of the major remaining challenges lies in modeling long-range motion-contained dynamic videos, where a naive extension of existing methods leads to severe memory explosion, temporal flickering, and failure to handle appearing or disappearing occlusions over time. To address these challenges, we propose a novel 4DGS framework characterized by an Anchor Relay-based Bidirectional Blending (ARBB) mechanism, named MoRel, which enables temporally consistent and memory-efficient modeling of long-range dynamic scenes. Our method progressively constructs locally canonical anchor spaces at key-frame time index and models inter-frame deformations at the anchor level, enhancing temporal coherence. By learning bidirectional deformations between KfA and adaptively blending them through learnable opacity control, our approach mitigates temporal discontinuities and flickering artifacts. We further introduce a Feature-variance-guided Hierarchical Densification (FHD) scheme that effectively densifies KfA's while keeping rendering quality, based on an assigned level of feature-variance. To effectively evaluate our model's capability to handle real-world long-range 4D motion, we newly compose long-range 4D motion-contained dataset, called SelfCap_{LR}. It has larger average dynamic motion magnitude, captured at spatially wider spaces, compared to previous dynamic video datasets. Overall, our MoRel achieves temporally coherent and flicker-free long-range 4D reconstruction while maintaining bounded memory usage, demonstrating both scalability and efficiency in dynamic Gaussian-based representations.