Research map / Learning theory and algorithms / Clustering
Learning the Clustering of Longitudinal Shape Data Sets into a Mixture of Independent or Branching Trajectories
Vianney Debavelaere, Stanley Durrleman, Stéphanie Allassonnière, Alzheimer’s Disease Neuroimaging Initiative
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add tempered (oscillating temperature) distribution to MCMC-SAEM cluster sampling (ablation) | MCMC-SAEM mixture · Simulated dancing man (100 subjects, branching) | classification error: 31.3% smaller after 500 iterations | Improved |
| replace geodesic regression + kmeans baseline with the mixture model, clusters differing in acceleration | Mixture model · Dancing men, acceleration clusters | cluster assignment accuracy: 51% -> 84% | Improved |
| replace baseline with the mixture model on easy simulated data (spatial difference only) | Mixture model · Simulated dancing man (100 subjects, branching) | cluster assignment accuracy: perfect classification by baseline vs 97% | No clear effect |
| replace baseline with the mixture model on 3D faces | Mixture model · 3D faces | cluster assignment accuracy: 88% -> 68.5% | Hurt |
| replace baseline with the mixture model on hippocampi | Mixture model · Hippocampi ADNI | cluster assignment accuracy: 52% -> 72% | Improved |
| constrain two templates to share first component (branching) instead of independent | Mixture model · RECIST scores | cluster separation: 28% / 72% size split -> 37% / 63% response split | Improved |
| select branching model over two independent trajectories using BIC | Mixture model · Simulated dancing man (100 subjects, branching) | BIC: 2.98% difference between BIC | Improved |
| select branching model over independent trajectories using BIC | Mixture model · Hippocampi ADNI | BIC: 2.92% difference between BIC | Improved |
| use log-likelihood ratio test instead of BIC for model selection | Mixture model · Simulated dancing man (100 subjects, branching) | model selection | Hurt |
Closest papers by meaning
- Learning the spatiotemporal variability in longitudinal shape data sets (IJCV 2020 · similarity 0.90)
- Learning Distributions of Shape Trajectories From Longitudinal Datasets: A Hierarchical Model on a Manifold of Diffeomorphisms (CVPR 2018 · similarity 0.85)
- Learning Segmentation from Point Trajectories (NeurIPS 2024 · similarity 0.81)
- Learned Trajectory Embedding for Subspace Clustering (CVPR 2024 · similarity 0.81)
- Improved Learning via k-DTW: A Novel Dissimilarity Measure for Curves (ICML 2025 · similarity 0.81)
- Clustering via Self-Supervised Diffusion (ICML 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
