Research map / Time series / Causal and generative time series
Uncertainty-aware tree height change regression
Max Gaber, Dimitri Gominski, Jaime Revenga, Stefan Oehmcke, Rasmus Fensholt, Martin Brandt
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 |
|---|---|---|---|
| train with tMSE loss instead of MSE (ablation) | DOFA · CHC test | RMSE: decreased by 4% | Improved |
| train with tMSE loss instead of MSE (ablation) | UNet · CHC test | RMSE: decreased by 9% | Improved |
| train with wMSE loss instead of MSE (ablation) | Prithvi · CHC test | RMSE: decreased by 8% | Improved |
| use uncertainty weighting or thresholding for weaker GFMs (ablation) | lower-performing GFMs · CHC test | RMSE | No clear effect |
| use tMSE instead of MSE for decrease detection (ablation) | ResNet-50 (frozen random) · CHC test (decrease) | F1: 0.51 -> 0.45 | Hurt |
| use tMSE instead of MSE for increase detection (ablation) | ResNet-50 (frozen random) · CHC test (increase) | F1: 0.75 -> 0.84 | Improved |
| use wMSE instead of tMSE for decrease detection (ablation) | Prithvi · CHC test (decrease) | F1: 0.58 -> 0.61 | Improved |
| use wMSE instead of MSE for decrease detection (ablation) | TerraMind · CHC test (decrease) | F1: 0.50 -> 0.44 | Hurt |
| use frozen GFM encoders instead of supervised UNet | GFMs (Prithvi, DOFA) · CHC test | RMSE: 5.65 -> 5.90 | Hurt |
| use ECHOSAT GEDI-trained dataset instead of GFMs/UNet | ECHOSAT · CHC test | RMSE: 5.65 -> 5.25 | Improved |
| use Scale-MAE encoder instead of random frozen ResNet-50 | Scale-MAE · CHC test | RMSE: 6.98 -> 7.83 | Hurt |
Closest papers by meaning
- Estimating Canopy Height at Scale (ICML 2024 · similarity 0.89)
- Capturing Temporal Dynamics in Large-Scale Canopy Tree Height Estimation (ICML 2025 · similarity 0.89)
- Open-Canopy: Towards Very High Resolution Forest Monitoring (CVPR 2025 · similarity 0.89)
- FoMo: Multi-Modal, Multi-Scale and Multi-Task Remote Sensing Foundation Models for Forest Monitoring (AAAI 2025 · similarity 0.85)
- The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generationf (CVPR 2025 · similarity 0.83)
- Multi-modal Learning for Geospatial Vegetation Forecasting (CVPR 2024 · similarity 0.83)
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.
