Research map / Video / Video segmentation and tracking
Spatio-Temporal Hilbert Maps for Continuous Occupancy Representation in Dynamic Environments
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Random instead of regular-grid support locations, 250 supports (ablation) | HF-STHM · Dataset 1, t*=0 | AUC: 0.95 -> 0.83 | Hurt |
| Random instead of regular-grid support locations, 500 supports (ablation) | HF-STHM · Dataset 1, t*=0 | AUC: 0.98 -> 0.88 | Hurt |
| Random instead of regular-grid support locations, 1000 supports (ablation) | HF-STHM · Dataset 1, t*=0 | AUC: 0.99 -> 0.94 | Hurt |
| Random instead of regular-grid supports, 5000 supports (ablation) | HF-STHM · Dataset 1, t*=0 | AUC: 0.99 -> 0.98 | Hurt |
| Embed distribution (mean and variance) instead of point estimate of future location (ablation) | HF-STHM · future maps, t*=1 to 5 | NLL decrease: 0.11, 0.22, 0.34, 0.83, 0.50, 1.36 | Helped |
| Use spatio-temporal HF-STHM instead of static Hilbert map | HF-STHM vs SHM · Dataset 2, t*=4 | NLL: 2.96 -> 0.16 | Helped |
| Use spatio-temporal HF-STHM instead of static Hilbert map | HF-STHM vs SHM · Dataset 1, t*=6 | NLL: 2.00 -> 1.68 | Helped |
| Use HF-STHM instead of SHM at present time | HF-STHM vs SHM · Dataset 1, t*=0 | NLL: 0.11 -> 0.12 | No clear effect |
| Use HF-STHM instead of dynamic occupancy grid map | HF-STHM vs DGM · Dataset 1, t*=6 | NLL: 34.54 -> 1.68 | Helped |
| Use HF-STHM instead of dynamic occupancy grid map | HF-STHM vs DGM · Dataset 2, t*=0 | NLL: 6.00 -> 0.09 | Helped |
| Compare with DGPOM for far-ahead prediction (t*>=2) | HF-STHM vs DGPOM · Dataset 1 future prediction | AUC | Helped |
| Increase prediction horizon t* | HF-STHM · Datasets 1 and 2 | NLL: 0.12 -> 1.68 | Hurt |
Closest papers by meaning
- Spatiotemporal Decoupling for Efficient Vision-Based Occupancy Forecasting (CVPR 2025 · similarity 0.84)
- Semi-Supervised Vision-Centric 3D Occupancy World Model for Autonomous Driving (ICLR 2025 · similarity 0.84)
- UnO: Unsupervised Occupancy Fields for Perception and Forecasting (CVPR 2024 · similarity 0.84)
- GTR: A General, Multi-View, and Dynamic Framework for Trajectory Representation Learning (ICML 2025 · similarity 0.83)
- Exploiting Continuous Motion Clues for Vision-Based Occupancy Prediction (AAAI 2025 · similarity 0.83)
- Driving in the Occupancy World: Vision-Centric 4D Occupancy Forecasting and Planning via World Models for Autonomous Driving (AAAI 2025 · 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.