SLAM++: Simultaneous Localisation and Mapping at the Level of Objects

By Renato F. Salas-Moreno, Richard A. Newcombe, Hauke Strasdat, Paul H. J. Kelly, Andrew J. Davison. 2013

998 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Geology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Geology

Citations in the corpus

Continuous Scene Representations for Embodied AI (2022, background): “There are object-based maps that encode object-level semantics and the relationships [3, 7, 42].”
The limits and potentials of deep learning for robotics (2018, neutral): “As an extension of reasoning over individual objects, inference over the geometry of the whole scene is important for robotic vision, and closely related to the problems of object-based mapping or object-based SLAM [19], [86], [98], [114].”
SUN3D: A Database of Big Spaces Reconstructed Using SfM and Object Labels (2013, methodology): “Discussion Although the generalized bundle adjustment is used here with manual annotation, it could potentially be extended to work with automatic object detection as well [3, 6].”
Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, neutral): “[217] propose a SLAM system based on the”
Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics (2017, background): “register the pose of 3D object models [8, 23, 53, 65] in order to index into a set of manipulation plans.”
Made to measure: Bespoke landmarks for 24-hour, all-weather localisation with a camera (2016, methodology): “used full objects in their framework rather than typical low-level features [25].”
Aligning Cyber Space With Physical World: A Comprehensive Survey on Embodied AI (2024, neutral): “Unlike previous datasets that explicitly specify target objects in questions, K-EQA [142] features complex questions with logical clauses and knowledge-related phrases, requiring prior knowledge to answer.”
Global Pose Estimation with an Attention-Based Recurrent Network (2018, background): “Several such solvers achieve excellent performance, for example, g2o [25], TSAM [8], Ceres, iSAM [20], SLAM++ [38], and recently [2] for optimization with semantic data association.”
DS-SLAM: A Semantic Visual SLAM towards Dynamic Environments (2018, background): “Therefore, combining these networks with SLAM could produce the semantic map, and thus improve the perception level of robots [5, 6].”
Fusion++: Volumetric Object-Level SLAM (2018, background): “Our pose-graph formulation is similar to that of [30].”
SemanticPaint (2015, background): “…et al. 2011; Anand et al. 2013; St¨uckler et al. 2013 meshes [Valentin et al. 2013], or voxel representations [Kim et al 2013a; Karpathy et al. 2013; Salas-Moreno et al. 2013; H¨ane et al. 2013 Methods have been proposed to break down meshes into semantic parts [Chen et al. 2009], perform dense…”
CosyPose: Consistent multi-view multi-object 6D pose estimation (2020, methodology): “This problem is known as object-level SLAM [26] where a depth-based object pose estimation method [27] is used to recognize objects from a database in individual images and estimate their poses.”
A Comprehensive Survey of Visual SLAM Algorithms (2022, methodology): “The SLAM++ algorithm [94] is an object-oriented SLAM algorithm that takes advantage of previously known scenes containing repeated objects and structures, such as a classroom.”
Kimera: an Open-Source Library for Real-Time Metric-Semantic Localization and Mapping (2019, background): “First, while Back-end Geometry Sema-ntics ORB-SLAM [22] mono g2o points DSO [23] mono g2o points VINS-mono [24] mono/IMU Ceres points VINS-Fusion [25] mono/Stereo/IMU Ceres points ROVIOLI [26] stereo/IMU EKF points ElasticFusion [18] RGB-D alternation surfels Voxblox [27] RGB-D [26] TSDF SLAM++ [16]”
A Robust 3D-2D Interactive Tool for Scene Segmentation and Annotation (2016, neutral): “, [29], [30], [31], [32], or part-based models, e.”
Fast Global Registration (2016, methodology): “This iterative process is typically based on variants of RANSAC [1, 19, 26, 29, 34] or pose clustering [8, 26, 35].”
Discrete-Continuous Depth Estimation from a Single Image (2014, background): “Therefore, they are now being integrated into 3D scene understanding methods that jointly detect, or segment objects while performing 3D reconstruction [6, 8, 23].”
Inferring 3D Object Pose in RGB-D Images (2015, background): “Finally, [27, 30] among many others, work on the same problem but either consider known instances of objects, or rely on user interaction.”
Revisiting active perception (2016, background): “Despite this success, however, the problem of localization and mapping in the absence of GPS, and when only visual sensing is available, still remains a challenge (Salas-Moreno et al. 2013), as we still lack notions of robust spatial representations.”