Surface creation on unstructured point sets using neural networks

Yumer, Mehmet Ersin; Kara, Levent Burak

doi:10.1016/j.cad.2012.03.002

Cited by 10 publications

(4 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The aforementioned methodology eliminates the requirement of complicated and time-consuming surface reconstruction methods, which often require sub-processes for noise elimination [27]. Normal vector estimation processes [28] could have also been implemented in order to automate the trimming step of the post-processing.…”

Section: Properly Treated Manualmentioning

confidence: 99%

Quality assurance for high-frequency mechanical impact (HFMI) treatment of welds using handheld 3D laser scanning technology

et al. 2014

View full text Add to dashboard Cite

The idea of using 3D point clouds obtained with the aid of a handheld 3D laser scanner for the quality assurance of high-frequency mechanical impact (HFMI) treatment is proposed and demonstrated in this paper. The effectiveness of impact treatments for extending the fatigue lives of welded structures has been demonstrated in numerous studies. Guidelines for the proper execution of impact treatments have been developed. A lack of suitable quality assurance (QA) procedures for accepting or rejecting the treatment after completion has been previously identified. In contrast with the existing QA procedures, which are based mainly on controlling inputs and visual inspection, a technology-based, quantitative methodology is developed in this paper. Five welded specimens were subjected to impact treatment at various levels to simulate under-, proper, and over-treatment. A handheld 3D laser scanner was then used to facilitate a point cloud-based method to determine the geometric parameters of the treated weld toe groove, which were then measured manually. The results show that the proposed methodology is successful in identifying the different treatment levels. This approach has a number of advantages over the existing QA methods, including the following: providing quantitative measures, ease of use, and archive-ability.

show abstract

Section: Properly Treated Manualmentioning

confidence: 99%

Quality assurance for high-frequency mechanical impact (HFMI) treatment of welds using handheld 3D laser scanning technology

et al. 2014

View full text Add to dashboard Cite

show abstract

“…In a complementary direction, Yumer et al . [24] developed a neural network for fitting a single NURBS patch to an unstructured point cloud. While the goal is similar to our spline-fitting network, it is not combined with a decomposition module that jointly learns how to express a shape with multiple patches covering different regions.…”

Section: Fittingmentioning

confidence: 99%

ParSeNet: A Parametric Surface Fitting Network for 3D Point Clouds

Sharma¹,

Liu²,

Maji³

et al. 2020

Preprint

View full text Add to dashboard Cite

We propose a novel, end-to-end trainable, deep network called ParSeNet that decomposes a 3D point cloud into parametric surface patches, including B-spline patches as well as basic geometric primitives.ParSeNet is trained on a large-scale dataset of man-made 3D shapes and captures high-level semantic priors for shape decomposition. It handles a much richer class of primitives than prior work, and allows us to represent surfaces with higher fidelity. It also produces repeatable and robust parametrizations of a surface compared to purely geometric approaches. We present extensive experiments to validate our approach against analytical and learning-based alternatives. ParseNet EditFig. 1: ParSeNet decomposes point clouds (top row) into collections of seamlessly assembled parametric surface patches including B-spline patches (bottom row). On the right, a shape is edited using the inferred parametrization.

show abstract

“…Since lapacian was used to depict geometric details of mesh surface model, it can make the filled mesh patch and the original mesh surface inosculate more naturally, but the filling result may lose some geometric details when the hole is too big. A neural network based B-spline surface construction has been proposed by Mehmet Ersin Yumer and Levent Burak Kara [12] , where an unstructured point set was projected into R2 by locally linear embedding (LLE) method in order to generate the parameterization. Then, a neural network was trained for mapping from the parameter space into 3D space.…”

Section: Introductionmentioning

confidence: 99%

A Novel Hole Filling Method based on the Hybrid PSO-BP Algorithm

Hong¹,

Wang²,

Zhang³

2016

IJCA

View full text Add to dashboard Cite

This paper presents a novel hole-filling algorithm in reverse engineering (RE) domain that can fill holes directly from the point clouds-a hybrid PSO-BP approach: Firstly, boundary of the hole is detected and feature points on the boundary are extracted. Secondly, a hole filling method based on the commercial reverse engineering software (Gemagic and Imageware) is employed to cover the hole with a rough mesh. Finally, a hybrid PSO-BP algorithm is exploited to refine the original mesh. The performance of the approach proposed has been evaluated by applying it to two different scattered point clouds from real-world scanned objects-a bucket of an excavator and a gear. The experimental results show that the suggested approach performs quite well, it is able to deal with highly accurate and extremely complicated data points. Besides, it can handle shapes with delicate details as well, the favorable fidelity and efficiency make it a promising candidate for many practical applications.

show abstract

Surface creation on unstructured point sets using neural networks

Cited by 10 publications

References 36 publications

Quality assurance for high-frequency mechanical impact (HFMI) treatment of welds using handheld 3D laser scanning technology

Quality assurance for high-frequency mechanical impact (HFMI) treatment of welds using handheld 3D laser scanning technology

ParSeNet: A Parametric Surface Fitting Network for 3D Point Clouds

A Novel Hole Filling Method based on the Hybrid PSO-BP Algorithm

Contact Info

Product

Resources

About