mlsanityData to 7 Oct 2026

MAGICIAN: Efficient Long-Term Planning with Imagined Gaussians for Active Mapping

arXiv 2603.22650 ↗First posted 23 Mar 2026cs.CV cs.ROcited by 5

Not on the map. Why?

Abstract (arXiv)

Active mapping aims to determine how an agent should move to efficiently reconstruct unknown environments. Most existing approaches rely on greedy next-best-view prediction, resulting in inefficient exploration and incomplete reconstruction. To address this, we introduce MAGICIAN, a novel long-term planning framework that maximizes accumulated surface coverage gain through Imagined Gaussians, a scene representation based on 3D Gaussian Splatting, derived from a pre-trained occupancy network with strong structural priors. This representation enables efficient coverage gain computation for any novel viewpoint via fast volumetric rendering, allowing its integration into a tree-search algorithm for long-horizon planning. We update Imagined Gaussians and refine the trajectory in a closed loop. Our method achieves state-of-the-art performance across indoor and outdoor benchmarks with varying action spaces, highlighting the advantage of long-term planning in active mapping.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 30 of its 49 references that are in the map. First posted on arXiv 23 Mar 2026.

Builds on

30 of 49 references.

References · 30

  1. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
  2. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  3. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,526
  4. 2025VGGT: Visual Geometry Grounded Transformercited by 1,475
  5. 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,354
  6. 20242D Gaussian Splatting for Geometrically Accurate Radiance Fieldscited by 1,061
  7. 2015ORB-SLAM: a Versatile and Accurate Monocular SLAM Systemcited by 983
  8. 2020Accelerating 3D Deep Learning with PyTorch3Dcited by 781
  9. 2023SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Renderingcited by 611
  10. 2022ActiveNeRF: Learning where to See with Uncertainty Estimationcited by 124
  11. 2024RaDe-GS: Rasterizing Depth in Gaussian Splattingcited by 110
  12. 2020Occupancy Anticipation for Efficient Exploration and Navigationcited by 104
18 more
  1. 2022Uncertainty Guided Policy for Active Robotic 3D Reconstruction using Neural Radiance Fieldscited by 53
  2. 2020FUEL: Fast UAV Exploration using Incremental Frontier Structure and Hierarchical Planningcited by 42
  3. 2025MILo: Mesh-In-the-Loop Gaussian Splatting for Detailed and Efficient Surface Reconstructioncited by 35
  4. 2024ActiveGS: Active Scene Reconstruction Using Gaussian Splattingcited by 34
  5. 2024ActiveSplat: High-Fidelity Scene Reconstruction through Active Gaussian Splattingcited by 32
  6. 2022Uncertainty-driven Planner for Exploration and Navigationcited by 31
  7. 2023Active Neural Mappingcited by 30
  8. 2024NARUTO: Neural Active Reconstruction from Uncertain Target Observationscited by 27
  9. 2019Supervised Learning of the Next-Best-View for 3D Object Reconstructioncited by 23
  10. 2020Fast Frontier-based Information-driven Autonomous Exploration with an MAVcited by 23
  11. 2025ActiveGAMER: Active GAussian Mapping through Efficient Renderingcited by 19
  12. 2023MACARONS: Mapping And Coverage Anticipation with RGB Online Self-Supervisioncited by 18
  13. 2022SCONE: Surface Coverage Optimization in Unknown Environments by Volumetric Integrationcited by 17
  14. 2017Submodular Trajectory Optimization for Aerial 3D Scanningcited by 16
  15. 2025NextBestPath: Efficient 3D Mapping of Unseen Environmentscited by 13
  16. 2020A Multi-Resolution Frontier-Based Planner for Autonomous 3D Explorationcited by 12
  17. 2025GLEAM: Learning Generalizable Exploration Policy for Active Mapping in Complex 3D Indoor Scenescited by 10
  18. 2025Understanding while Exploring: Semantics-driven Active Mappingcited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026R2S-EGO: Dual-Proxy Refinement for Sparse-Capture Real-to-Simcited by 1
  2. 2026Remember to be Curious: Episodic Context and Persistent Worlds for 3D Explorationcited by 0
  3. 2026Diagnosing and Dynamically Filtering Occupancy World Models for Active Mappingcited by 0
  4. 2026Rethinking Learned Occupancy in Autonomous Active Mapping with Observation-Gated Filteringcited by 0
  5. 2026Matisse: Evidence-Space Reasoning for Active 3D Reconstructioncited by 0