# Dyna3DGR: 4D Cardiac Motion Tracking with Dynamic 3D Gaussian Representation

arXiv: https://arxiv.org/abs/2507.16608 · HTML: https://mlsanity.com/paper/2507.16608/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 28 references that are in the map.
- First posted on arXiv 22 Jul 2025.

## Builds on (most cited first, 18 of 18)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [VoxelMorph: A Learning Framework for Deformable Medical Image Registration](https://mlsanity.com/paper/1809.05231/) (2018)
- [SC-GS: Sparse-Controlled Gaussian Splatting for Editable Dynamic Scenes](https://mlsanity.com/paper/2312.14937/) (2023)
- [Learning a Probabilistic Model for Diffeomorphic Registration](https://mlsanity.com/paper/1812.07460/) (2018)
- [DiffuseMorph: Unsupervised Deformable Image Registration Using Diffusion Model](https://mlsanity.com/paper/2112.05149/) (2021)
- [Correlation-aware Coarse-to-fine MLPs for Deformable Medical Image Registration](https://mlsanity.com/paper/2406.00123/) (2024)
- [DeepTag: An Unsupervised Deep Learning Method for Motion Tracking on Cardiac Tagging Magnetic Resonance Images](https://mlsanity.com/paper/2103.02772/) (2021)
- [Region-specific Diffeomorphic Metric Mapping](https://mlsanity.com/paper/1906.00139/) (2019)
- [Biomechanics-informed Neural Networks for Myocardial Motion Tracking in MRI](https://mlsanity.com/paper/2006.04725/) (2020)
- [Generative Myocardial Motion Tracking via Latent Space Exploration with Biomechanics-informed Prior](https://mlsanity.com/paper/2206.03830/) (2022)
- [FOAL: Fast Online Adaptive Learning for Cardiac Motion Estimation](https://mlsanity.com/paper/2003.04492/) (2020)
- [FSDiffReg: Feature-wise and Score-wise Diffusion-guided Unsupervised Deformable Image Registration for Cardiac Images](https://mlsanity.com/paper/2307.12035/) (2023)
- [Explainable cardiac pathology classification on cine MRI with motion characterization by semi-supervised learning of apparent flow](https://mlsanity.com/paper/1811.03433/) (2018)
- [4D Myocardium Reconstruction with Decoupled Motion and Shape Model](https://mlsanity.com/paper/2308.14083/) (2023)
- [DeepMesh: Mesh-based Cardiac Motion Tracking using Deep Learning](https://mlsanity.com/paper/2309.14306/) (2023)
- [Unsupervised Landmark Detection Based Spatiotemporal Motion Estimation for 4D Dynamic Medical Images](https://mlsanity.com/paper/2109.14805/) (2021)
- [3DGR-CAR: Coronary artery reconstruction from ultra-sparse 2D X-ray views with a 3D Gaussians representation](https://mlsanity.com/paper/2410.00404/) (2024)
- [Bidirectional Recurrence for Cardiac Motion Tracking with Gaussian Process Latent Coding](https://mlsanity.com/paper/2410.20752/) (2024)

## Built on by (most cited first, 1 of 1)

- [Anatomy-Aligned Surface Field Learning for Myocardial Reconstruction from Sparse Short-Axis Cine MRI](https://mlsanity.com/paper/2609.29825/) (2026)

## Abstract (arXiv)

Accurate analysis of cardiac motion is crucial for evaluating cardiac function. While dynamic cardiac magnetic resonance imaging (CMR) can capture detailed tissue motion throughout the cardiac cycle, the fine-grained 4D cardiac motion tracking remains challenging due to the homogeneous nature of myocardial tissue and the lack of distinctive features. Existing approaches can be broadly categorized into image based and representation-based, each with its limitations. Image-based methods, including both raditional and deep learning-based registration approaches, either struggle with topological consistency or rely heavily on extensive training data. Representation-based methods, while promising, often suffer from loss of image-level details. To address these limitations, we propose Dynamic 3D Gaussian Representation (Dyna3DGR), a novel framework that combines explicit 3D Gaussian representation with implicit neural motion field modeling. Our method simultaneously optimizes cardiac structure and motion in a self-supervised manner, eliminating the need for extensive training data or point-to-point correspondences. Through differentiable volumetric rendering, Dyna3DGR efficiently bridges continuous motion representation with image-space alignment while preserving both topological and temporal consistency. Comprehensive evaluations on the ACDC dataset demonstrate that our approach surpasses state-of-the-art deep learning-based diffeomorphic registration methods in tracking accuracy. The code will be available in https://github.com/windrise/Dyna3DGR.