Point cloud data filtering and downsampling using growing neural gas

Orts-Escolano, Sergio; Morell, Vicente; García-Rodríguez, José; Cazorla, Miguel

doi:10.1109/ijcnn.2013.6706719

Cited by 40 publications

(29 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This procedure is interesting when the purpose of the method that uses these data needs to have information about the whole space, such as rendering or surface reconstruction. Moreover, GNG has demonstrated noise reduction capabilities in [18]. The normal-based sampling returns higher density in those parts with higher shape variation.…”

Section: Evaluation Of Downsampling Methodsmentioning

confidence: 99%

“…The GNG-based sampling method also includes other benefits, such as the capability to filter noise in the 3D point cloud preserving the topology of the input space in the output representation [18,17].…”

Section: Sampling Gng-basedmentioning

confidence: 99%

“…Orts-Escolano et al presented in [18] a point cloud sampling based on artificial neural networks. Concretely, they proposed the use of Growing Neural Gas to obtain a simplification of the point cloud where each neuron of the final GNG represents one point.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Evaluation of sampling method effects in 3D non-rigid registration

Saval-Calvo

Azorín-López

Fuster-Guilló

et al. 2016

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

Since the beginning of 3D computer vision problems, the use of techniques to reduce the data to make it treatable preserving the important aspects of the scene has been necessary. Currently, with the new low-cost RGB-D sensors, which provide a stream of color and 3D data of approximately 30 frames per second, this is getting more relevance. Many applications make use of these sensors, and need a preprocessing to downsample the data in order to either reduce the processing time or improve the data (e.g. reducing noise or enhancing the important features). In this paper, we present a comparison of different downsampling techniques which are based on different principles. Concretely, five different downsampling methods are included: a bilinear-based method, a normal-based, a color-based, a combination of the normal and colorbased samplings, and a Growing Neural Gas (GNG) based approach. For the comparison, two different models have been used acquired with the Blensor software. Moreover, to evaluate the effect of the downsampling in a real application, a 3D non-rigid registration is performed with the data sampled. From the experimentation we can conclude that depending on the purpose of the application some kernels of the sampling methods can improve drastically the results. Bilinear and GNG-based methods provide homogeneous point clouds, but color-based and normal-based provide datasets with higher density of points in areas with specific features. In the non-rigid application, This study was supported in part by the University of Alicante and Spanish government under grants GRE11-01 and DPI2013-40534-R. if a color-based sampled point cloud is used, it is possible to properly register two datasets for cases where intensity data is relevant in the model, and outperform the results if only a homogeneous sampling is used.

show abstract

Section: Evaluation Of Downsampling Methodsmentioning

confidence: 99%

Section: Sampling Gng-basedmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of sampling method effects in 3D non-rigid registration

Saval-Calvo

Azorín-López

Fuster-Guilló

et al. 2016

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

show abstract

“…For each individual voxel, the points that lie within it are down-sampled with respect to their centroid. This approach has a few drawbacks: (i) it requires a slightly longer processing time as opposed to using the voxel center; (ii) it is sensitive to noisy input spaces; and (iii) it does not represent the underlying surface accurately [23].…”

Section: Voxel Gridmentioning

confidence: 99%

Frame Filtering and Skipping for Point Cloud Data Video Transmission

Moreno¹,

Li²

2017

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

“…Since common processing steps in 3D computer vision problems as 3D keypoint detection and 3D feature extraction algorithms use surface normal information [16,17,18]. As it was demonstrated in [19], the reduced representation created by the GNG algorithm has benefits for removing noise from data provided by low-cost sensors, improving the performance in some computer vision problems as 3D scene recognition. Normal information is included during the learning step of the GNG, and therefore, we are able to produce a reduced representation of the surface normal information.…”

Section: Preserving Surface Normal Informationmentioning

confidence: 99%

3D Surface Reconstruction of Noisy Point Clouds Using Growing Neural Gas: 3D Object/Scene Reconstruction

Orts-Escolano

García-Rodríguez

Morell

et al. 2015

Neural Process Lett

Self Cite

View full text Add to dashboard Cite

With the advent of low-cost 3D sensors and 3D printers, scene and object 3D surface reconstruction has become an important research topic in the last years. In this work, we propose an automatic (unsupervised) method for 3D surface reconstruction from raw unorganized point clouds acquired using low-cost 3D sensors. We have modified the Growing Neural Gas (GNG) network, which is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation, to perform 3D surface reconstruction of different real-world objects and scenes. Some improvements have been made on the original algorithm considering colour and surface normal information of input data during the learning stage and creating complete triangular meshes instead of basic wire-frame representations. The proposed method is able to successfully create 3D faces online, whereas existing 3D reconstruction methods based on Self-Organizing Maps (SOMs) required postprocessing steps to close gaps and holes produced during the 3D reconstruction process. A set of quantitative and qualitative experiments were carried out to validate the proposed method. The method has been implemented and tested on real data, and has been found to be effective at reconstructing noisy point clouds obtained using low-cost 3D sensors.

show abstract

Point cloud data filtering and downsampling using growing neural gas

Cited by 40 publications

References 21 publications

Evaluation of sampling method effects in 3D non-rigid registration

Evaluation of sampling method effects in 3D non-rigid registration

Frame Filtering and Skipping for Point Cloud Data Video Transmission

3D Surface Reconstruction of Noisy Point Clouds Using Growing Neural Gas: 3D Object/Scene Reconstruction

Contact Info

Product

Resources

About