Conditional Latent Diffusion Model with Fourier-based Motion Modelling for Virtual Population Synthesis
2026-06-02 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors developed a new computer model called 4D F-MeshLDM to create realistic moving 3D heart shapes over time for virtual medical trials. Their method uses math tricks like Fourier series to capture the heartbeat’s repeating motion and a special neural network to generate accurate heart shapes based on patient data. Tested on large real-world heart data, their model made better and more realistic heart motion shapes than previous methods. This approach could help simulate heart device tests more reliably without needing real patients.
in-silico trialsgenerative model3D+t meshlatent spaceFourier seriesdiffusion modelvariational autoencoder (VAE)UK Biobankcardiac motionclinical covariates
Authors
Shaokun Lan, Haoran Dou, Jinghan Huang, Arezoo Zakeri, Fengming Lin, Zherui Zhou, Jinming Duan, Alejandro F. Frangi
Abstract
In-silico trials of medical devices require the generation of virtual populations of anatomies. In cardiovascular applications, virtual anatomy is typically represented as a 3D+t mesh sampled from a generative model. However, most existing mesh generators focus on static anatomy, while sequence models often lack explicit periodicity. To this end, we propose 4D F-MeshLDM, a conditional generative framework comprising a convolutional mesh VAE to encode meshes, a structural latent space that parameterises motion using a truncated Fourier series, and a diffusion prior that learns the latent distribution over Fourier coefficient tokens. By conditioning the diffusion process on clinical covariates via affine modulation, we enable controllable synthesis. Sampling tokens and performing inverse Fourier synthesis yield cycle-consistent latent trajectories, which can be decoded into 3D+t cardiac mesh sequences. Experiments on 5,000 UK Biobank subjects demonstrate that 4D F-MeshLDM outperforms state-of-the-art baselines in anatomical fidelity and achieves near-zero cycle closure error. Furthermore, the generated cohorts accurately preserve clinical functional indices, highlighting the potential of our framework for reliable in-silico cardiac trials.