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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Learning a powerful representation from point clouds is a fundamental andchallenging problem in the field of computer vision. Different from imageswhere RGB pixels are stored in the regular grid, for point clouds, theunderlying semantic and structural information of point clouds is the spatiallayout of the points. Moreover, the properties of challenging in-context andbackground noise pose more challenges to point cloud analysis. One assumptionis that the poor performance of the classification model can be attributed tothe indistinguishable embedding feature that impedes the search for the optimalclassifier. This work offers a new strategy for learning powerfulrepresentations via a contrastive learning approach that can be embedded intoany point cloud classification network. First, we propose a supervisedcontrastive classification method to implement embedding feature distributionrefinement by improving the intra-class compactness and inter-classseparability. Second, to solve the confusion problem caused by smallinter-class compactness and inter-class separability. Second, to solve theconfusion problem caused by small inter-class variations between somesimilar-looking categories, we propose a confusion-prone class mining strategyto alleviate the confusion effect. Finally, considering that outliers of thesample clusters in the embedding space may cause performance degradation, wedesign an entropy-aware attention module with information entropy theory toidentify the outlier cases and the unstable samples by measuring theuncertainty of predicted probability. The results of extensive experimentsdemonstrate that our method outperforms the state-of-the-art approaches byachieving 82.9% accuracy on the real-world ScanObjectNN dataset and substantialperformance gains up to 2.9% in DCGNN, 3.1% in PointNet++, and 2.4% in GBNet.

Contrastive Embedding Distribution Refinement and Entropy-Aware Attention for 3D Point Cloud Classification