Changes in Real Time: Online Scene Change Detection with Multi-View Fusion
Not on the map. Why?
Abstract (arXiv)
Online Scene Change Detection (SCD) is an extremely challenging problem that requires an agent to detect relevant changes on the fly while observing the scene from unconstrained viewpoints. Existing online SCD methods are significantly less accurate than offline approaches. We present the first online SCD approach that is pose-agnostic, label-free, and ensures multi-view consistency, while operating at over 10 FPS and achieving new state-of-the-art performance, surpassing even the best offline approaches. Our method introduces a new self-supervised fusion loss to infer scene changes from multiple cues and observations, PnP-based fast pose estimation against the reference scene, and a fast change-guided update strategy for the 3D Gaussian Splatting scene representation. Extensive experiments on complex real-world datasets demonstrate that our approach outperforms both online and offline baselines.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 40 of its 54 references that are in the map. First posted on arXiv 15 Nov 2025.
Builds on
40 of 54 references.
References · 40
- 2023Segment Anythingcited by 8,234
- 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
- 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
- 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,723
- 2024SAM 2: Segment Anything in Images and Videoscited by 2,863
- 20234D Gaussian Splatting for Real-Time Dynamic Scene Renderingcited by 1,157
- 2018From Coarse to Fine: Robust Hierarchical Localization at Large Scalecited by 559
- 2018Fully Convolutional Siamese Networks for Change Detectioncited by 306
- 2024XFeat: Accelerated Features for Lightweight Image Matchingcited by 113
- 2024Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitivescited by 95
- 2024RTAB-Map as an Open-Source Lidar and Visual SLAM Library for Large-Scale and Long-Term Online Operationcited by 62
- 2023PAD: A Dataset and Benchmark for Pose-agnostic Anomaly Detectioncited by 46
28 more
- 2018Weakly Supervised Silhouette-based Semantic Scene Change Detectioncited by 44
- 20243DGS-LM: Faster Gaussian-Splatting Optimization with Levenberg-Marquardtcited by 43
- 2021Panoptic Multi-TSDFs: a Flexible Representation for Online Multi-resolution Volumetric Mapping and Long-term Dynamic Scene Consistencycited by 39
- 2024Khronos: A Unified Approach for Spatio-Temporal Metric-Semantic SLAM in Dynamic Environmentscited by 36
- 2025On-the-fly Reconstruction for Large-Scale Novel View Synthesis from Unposed Imagescited by 36
- 2021DR-TANet: Dynamic Receptive Temporal Attention Network for Street Scene Change Detectioncited by 32
- 2020Hierarchical Paired Channel Fusion Network for Street Scene Change Detectioncited by 31
- 2021ChangeSim: Towards End-to-End Online Scene Change Detection in Industrial Indoor Environmentscited by 28
- 2024SplatPose & Detect: Pose-Agnostic 3D Anomaly Detectioncited by 25
- 2023CLNeRF: Continual Learning Meets NeRFcited by 24
- 2025LongSplat: Robust Unposed 3D Gaussian Splatting for Casual Long Videoscited by 24
- 2023The Change You Want to See (Now in 3D)cited by 17
- 2024Towards Generalizable Scene Change Detectioncited by 17
- 2023Living Scenes: Multi-object Relocalization and Reconstruction in Changing 3D Environmentscited by 15
- 20243DGS-CD: 3D Gaussian Splatting-based Change Detection for Physical Object Rearrangementcited by 15
- 2024Zero-Shot Scene Change Detectioncited by 13
- 2024ZeroSCD: Zero-Shot Street Scene Change Detectioncited by 12
- 2024Robust Scene Change Detection Using Visual Foundation Models and Cross-Attention Mechanismscited by 12
- 2024Multi-View Pose-Agnostic Change Localization with Zero Labelscited by 12
- 2024SplatPose+: Real-time Image-Based Pose-Agnostic 3D Anomaly Detectioncited by 11
- 2025GaussianUpdate: Continual 3D Gaussian Splatting Update for Changing Environmentscited by 11
- 2025CL-Splats: Continual Learning of Gaussian Splatting with Local Optimizationcited by 10
- 2025EMPLACE: Self-Supervised Urban Scene Change Detectioncited by 7
- 2023POV-SLAM: Probabilistic Object-Aware Variational SLAM in Semi-Static Environmentscited by 6
- 2025Gaussian Difference: Find Any Change Instance in 3D Scenescited by 5
- 2025LT-Gaussian: Long-Term Map Update Using 3D Gaussian Splatting for Autonomous Drivingcited by 5
- 2022PlaneSDF-based Change Detection for Long-term Dense Mappingcited by 3
- 2025Gaussian Mapping for Evolving Scenescited by 3
Built on by
Citing papers per year: 2026: 3.
3 papers.