Fast Global Registration

By Qian-Yi Zhou, Jaesik Park, Vladlen Koltun. European Conference on Computer Vision, 2016

1,124 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/7f7948942a3064c053d7b85f178994eddaabff87

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 Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Deep Closest Point: Learning Representations for Point Cloud Registration (2019, methodology): “Although the learned representations are task-dependent, DCP-v2 exhibits smaller error than others except for FGR, including the learning-based method PointNetLK.”
3DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructions (2016, neutral): “[43] 51.”
Graduated Non-Convexity for Robust Spatial Perception: From Non-Minimal Solvers to Global Outlier Rejection (2019, methodology): “For instance, [35] focuses on a specific problem (point cloud registration) where (9) can be solved in closed form [3], [36], while [11] focuses on a Markov Random Field formulation for which global solvers and heuristics exist [47].”
PointDSC: Robust Point Cloud Registration using Deep Spatial Consistency (2021, methodology): “We first select four representative traditional methods: FGR [77], SM [36], RANSAC [24], and GC-RANSAC [5], as well as the state-of-the-art geometrybased method TEASER [66].”
Deep Global Registration (2020, methodology): “…} and F y = { f y 1 , ..., f y Nx } of two 3D scans, we use the nearest neighbor in the feature space to generate a set of putative correspondences or matches M = { ( i, This procedure is deterministic and can be hand-crafted to ﬁlter out noisy correspondences with ratio or reciprocity tests [50].”
Vision-based robotic grasping from object localization, object pose estimation to grasp estimation for parallel grippers: a review (2020, background): “1 3 registration methods (Mellado et al. 2014; Yang et al. 2015; Zhou et al. 2016) exist which afford large variations of initial poses and are robust with large noise.”
PRNet: Self-Supervised Learning for Partial-to-Partial Registration (2019, neutral): “We compare our model to ICP, Go-ICP [14], Fast Global Registration (FGR) [61], and DCP [2].”
A General and Adaptive Robust Loss Function (2017, methodology): “In Figure 5 we set α to a wide range of values and report errors for each setting, using the same evaluation of [40].”
DeepIM: Deep Iterative Matching for 6D Pose Estimation (2018, background): “Global registration methods (Mellado et al. 2014; Theiler et al. 2015; Zhou et al. 2016; Yang et al. 2016) solve a more challenging problem by not assuming an initial pose estimate.”
A comprehensive survey on point cloud registration (2021, methodology): “FGR [119] is selected for the classic optimization-based algorithm.”
A self-supervised learning system for object detection using physics simulation and multi-view pose estimation (2017, methodology): “Nevertheless, in application domains, such as those studied by the Amazon Picking Challenge [2], which involve varying light conditions and cluttered scenes, it has been shown [5] that CNN-based segmentation [10], [15] followed by point cloud registration with 3D models [23], [24], [25] is an effect”
Fast and Robust Iterative Closest Point (2020, methodology): “We further test the methods on 25 pairs of partially over-lapping point clouds constructed from ﬁve models in [30], with ﬁve pairs for each model.”
Colored Point Cloud Registration Revisited (2017, background): “Such initial alignment can be obtained via global registration algorithms [18, 43, 47].”
Open3D: A Modern Library for 3D Data Processing (2018, background): “Open3D has been in development since 2015 and has been used in a number of published research projects [22, 13, 15, 12].”
RPM-Net: Robust Point Matching Using Learned Features (2020, methodology): “Note that ICP and FGR are executed on CPU and the remaining algorithms on a GPU.”