Joint Layout Estimation and Global Multi-view Registration for Indoor Reconstruction

@article{Lee2017JointLE,
  title={Joint Layout Estimation and Global Multi-view Registration for Indoor Reconstruction},
  author={Jeong-Kyun Lee and Jae-Won Yea and M. Park and Kuk-jin Yoon},
  journal={2017 IEEE International Conference on Computer Vision (ICCV)},
  year={2017},
  pages={162-171}
}
In this paper, we propose a novel method to jointly solve scene layout estimation and global registration problems for accurate indoor 3D reconstruction. Given a sequence of range data, we first build a set of scene fragments using KinectFusion and register them through pose graph optimization. Afterwards, we alternate between layout estimation and layout-based global registration processes in iterative fashion to complement each other. We extract the scene layout through hierarchical… 

Figures and Tables from this paper

Floorplan-Jigsaw: Jointly Estimating Scene Layout and Aligning Partial Scans
TLDR
Without the requirement of feature matching, the proposed method enables some novel applications ranging from large or featureless scene reconstruction and modeling from sparse input, and quantitatively and qualitatively on real and synthetic scenes of various sizes and complexities.
Floorplan Priors for Joint Camera Pose and Room Layout Estimation
TLDR
This work exploits the 3D "box" structure of a typical room layout that meets the Manhattan World property to reconstruct large or featureless scenes from a set of partial reconstructions due to camera tracking interruptions when scanning a large orfeatureless scene.
Real-time dense 3D reconstruction and camera tracking via embedded planes representation
TLDR
A novel approach for robust plane matching and real-time RGB-D fusion based on the representation of plane parameter space, which explores the holistic topology of all relevant planes and adopts the low-dimensionality parameter space representation.
MVLayoutNet: 3D layout reconstruction with multi-view panoramas
TLDR
MVLayoutNet is presented, an end-to-end network for holistic 3D reconstruction from multi-view panoramas that jointly train a layout module to produce an initial layout and a novel MVS module to obtain accurate layout geometry.
Real-Time Global Registration for Globally Consistent RGB-D SLAM
TLDR
Experiments show that the proposed algorithm achieves globally consistent pose estimation in real-time via CPU computing, and owns comparable accuracy as state-of-the-art that use GPU computing, enabling the practical usage of globally consistent RGB-D SLAM on highly computationally constrained devices.
Monocular Object and Plane SLAM in Structured Environments
TLDR
Experiments show that the monocular simultaneous localization and mapping algorithm can improve camera localization accuracy compared to state-of-the-art SLAM, especially when there is no loop closure, and also generate dense maps robustly in many structured environments.
Semantic Object and Plane SLAM for RGB-D Cameras
TLDR
This work takes advantage of state of the art object detectors and a robust ICP method to build a map made up of objects providing semantic information as well as extra constraints to original point-based SLAM system, and introduces plane landmarks into SLAM framework to reduce drift.
Real-Time Calibration and Registration Method for Indoor Scene with Joint Depth and Color Camera
TLDR
A novel registration method that takes advantages of RGB-D camera to obtain the depth information in real-time, and a binocular system using the Time of Flight (ToF) camera and a commercial color camera is constructed to realize the three-dimensional registration technique.
Towards Automated Floorplan Generation
TLDR
This paper proposes a pipeline for generating a 2D floorplan using depth cameras, using an existing approach to recovering the camera motion trajectories from the depth and RGB sequences and examines the comparative accuracies of two ground plane estimation methods for each of the above mentioned approaches to rectilinear abstraction.
Flat2Layout: Flat Representation for Estimating Layout of General Room Types
TLDR
A dynamic programming based postprocessing is employed to decode the estimated flat output from the deep model into the final room layout, which achieves state-of-the-art performance on existing room layout benchmark.
...
...

References

SHOWING 1-10 OF 53 REFERENCES
Joint 3D Estimation of Objects and Scene Layout
TLDR
A novel generative model is proposed that is able to reason jointly about the 3D scene layout as well as the3D location and orientation of objects in the scene and significantly increase the performance of state-of-the-art object detectors in their ability to estimate object orientation.
Online Structure Analysis for Real-Time Indoor Scene Reconstruction
TLDR
This work proposes a real-time approach for indoor scene reconstruction that is capable of producing a ready-to-use 3D geometric model even while the user is still scanning the environment with a consumer depth camera, via an online structure analysis on the dynamic, incomplete data.
Free your Camera: 3D Indoor Scene Understanding from Arbitrary Camera Motion
TLDR
This work proposes an effective probabilistic formulation that allows us to generate, evaluate and optimize layout hypotheses by integrating new image evidence as the observer moves, and demonstrates that this formulation reaches near-real-time computation time and outperforms state-of-the-art methods while operating in significantly less constrained conditions.
Exploring High-Level Plane Primitives for Indoor 3D Reconstruction with a Hand-held RGB-D Camera
TLDR
This work proposes to extract high level primitives---planes---from an RGB-D camera, in addition to low level image features, to better constrain the problem and help improve indoor 3D reconstruction and demonstrates with real datasets that the method with plane constraints achieves more accurate and more appealing results comparing with other state-of-the-art scene reconstruction algorithms.
Hierarchical Multiview Rigid Registration
TLDR
The proposed hierarchical approach is accurate, efficient and robust for initial view states that are not well posed and improves the accuracy and robustness of registration by simultaneously considering transformation and registration errors.
Towards a General Multi-View Registration Technique
TLDR
An algorithm is presented that reduces significantly the level of the registration errors between all pairs in a set of range views and improves the calibrated registrations and the quality of the integrated model for complex multi-part objects.
Indoor scene structure analysis for single image depth estimation
TLDR
This work introduces a hierarchical representation of the scene, which models local depth jointly with mid-level and global scene structures and formulate single image depth estimation as inference in a graphical model whose edges let us encode the interactions within and across the different layers of the authors' hierarchy.
Robust reconstruction of indoor scenes
TLDR
An approach to indoor scene reconstruction from RGB-D video to combine geometric registration of scene fragments with robust global optimization based on line processes that substantially increases the accuracy of reconstructed scene models.
Dense planar SLAM
TLDR
This work presents an efficient new real-time approach which densely maps an environment using bounded planes and surfels extracted from depth images (like those produced by RGB-D sensors or dense multi-view stereo reconstruction) to take advantage of the planarity of many parts of real-world scenes.
Deformation-based loop closure for large scale dense RGB-D SLAM
TLDR
An incremental method for deformation graph construction is derived, allowing multi-million point maps to be captured over hundreds of metres in real-time, allowing for highly accurate dense maps over large scale trajectories that are both locally and globally consistent.
...
...