Connect-and-Slice: An Hybrid Approach for Reconstructing 3D Objects
read more
Citations
Inferring CAD Modeling Sequences Using Zone Graphs
Floorplan generation from 3D point clouds: A space partitioning approach
Structure-aware indoor scene reconstruction via two levels of abstraction
City3D: Large-Scale Building Reconstruction from Airborne LiDAR Point Clouds
Urban Scene LOD Vectorized Modeling From Photogrammetry Meshes
References
Surface simplification using quadric error metrics
Poisson surface reconstruction
Efficient RANSAC for Point-Cloud Shape Detection
Screened poisson surface reconstruction
On the shape of a set of points in the plane
Related Papers (5)
Frequently Asked Questions (15)
Q2. What have the authors stated for future works in "Connect-and-slice: an hybrid approach for reconstructing 3d objects" ?
In future work the authors would like to investigate on its automatic selection. The authors also wish to understand the hierarchical relationships between primitives in order to detect and exploit high order structural information as symmetry.
Q3. How does the algorithm perform on more challenging datasets?
On more challenging datasets where hundred primitives are necessary to decently approximate the objects as Rubbish bin, their algorithm performs better in terms of visual quality, output complexity and running time.
Q4. What are the main problems of connectivity-based methods?
Despite being fast, connectivity-based methods suffer from a lack of robustness to defect-laden data, in particular to over- and under-detection of primitives and erroneous connections between primitives.
Q5. What is the simplest way to penalize the presence of facets?
• Structural constraint imposes the structural facets to be active, i.e. part of the output surface (Eq. 4):xi = 1, ∀i ∈ Fs (4)where Fs corresponds to the set of structural facets.
Q6. What is the role of the soft-connectivity relationship?
Parameter specifying the soft-connectivity relationship plays an important role in controlling how far the connected primitives can be located from each others.
Q7. What types of acquisition systems have been used to generate the datasets?
Different types of acquisition systems have been used to generate the datasets, including Laser, e.g. Euler and Hand, multi-view stereo, e.g. Cottage and Building block, and Kinect, e.g. Rubbish bin and Couch.
Q8. What is the key ingredient to solve this problem?
Their approach proposes two key ingredients to solve this issue: a new light and spatially-adaptive partitioning data-structure and a preliminary connectivity analysis that reduces the number of primitives to be processed during slicing operations.
Q9. What is the slicing domain of the 3D bounding box?
the 3D bounding box is divided into polyhedra by inserting one per one each slicing domain in an arbitrary order: the first slicing domain splits the 3D bounding box into two polyhedra, the second slicing domain typically splits the two polyhedra into four polyhedra, etc.
Q10. What is the way to detect geometric primitives from 3D data?
Analyzing a connectivity graph to detect and link points intersecting plane triples [8, 29, 33] usually works well when the correct connectivity between primitives can be recovered.
Q11. What is the way to represent a scene?
One prefers representing such scenes by more compact and structure-aware Computer-Aided Design style models, i.e. with concise polygon meshes in which each facet corresponds to a large polygon [4].
Q12. What is the simplest way to reduce the number of facets on the 3D bounding?
As input points have often missing parts on their 3D bounding box (see for instance Church and Face in Figure 7), the authors relax the watertight constraint on edges lying on the 3D bounding box.
Q13. What is the main idea of the algorithm?
The algorithm is built on several key technical ingredients that allows us to operate on an efficient and compact partitioning data-structure.
Q14. How do the authors build the partitioning data-structure in Section 5?
The authors then build the partitioning data-structure in Section 5 by slicing the spatially-close unprocessed primitives while embedding the structurally-valid components found in the previous step.
Q15. How do the authors determine if two primitives are mutual neighbors?
When detected from point clouds, two primitives are said strongly-connected if at least two inlier points fitted each to one of the two primitives are mutual neighbors in the k-nearest neighbor graph of the input points.