English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

3D point clouds have attracted increasing attention in architecture,engineering, and construction due to their high-quality object representationand efficient acquisition methods. Consequently, many point cloud featuredetection methods have been proposed in the literature to automate someworkflows, such as their classification or part segmentation. Nevertheless, theperformance of point cloud automated systems significantly lags behind theirimage counterparts. While part of this failure stems from the irregularity,unstructuredness, and disorder of point clouds, which makes the task of pointcloud feature detection significantly more challenging than the image one, weargue that a lack of inspiration from the image domain might be the primarycause of such a gap. Indeed, given the overwhelming success of ConvolutionalNeural Networks (CNNs) in image feature detection, it seems reasonable todesign their point cloud counterparts, but none of the proposed approachesclosely resembles them. Specifically, even though many approaches generalisethe convolution operation in point clouds, they fail to emulate the CNNsmultiple-feature detection and pooling operations. For this reason, we proposea graph convolution-based unit, dubbed Shrinking unit, that can be stackedvertically and horizontally for the design of CNN-like 3D point cloud featureextractors. Given that self, local and global correlations between points in apoint cloud convey crucial spatial geometric information, we also leverage themduring the feature extraction process. We evaluate our proposal by designing afeature extractor model for the ModelNet-10 benchmark dataset and achieve90.64% classification accuracy, demonstrating that our innovative idea iseffective. Our code is available at github.com/albertotamajo/Shrinking-unit.

Shrinking unit: a Graph Convolution-Based Unit for CNN-like 3D Point Cloud Feature Extractors