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/

Zendy Plus

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

Zendy Tools

Zendy Open

In this paper, we propose a novel learning-based pipeline for partiallyoverlapping 3D point cloud registration. The proposed model includes aniterative distance-aware similarity matrix convolution module to incorporateinformation from both the feature and Euclidean space into the pairwise pointmatching process. These convolution layers learn to match points based on jointinformation of the entire geometric features and Euclidean offset for eachpoint pair, overcoming the disadvantage of matching by simply taking the innerproduct of feature vectors. Furthermore, a two-stage learnable pointelimination technique is presented to improve computational efficiency andreduce false positive correspondence pairs. A novel mutual-supervision loss isproposed to train the model without extra annotations of keypoints. Thepipeline can be easily integrated with both traditional (e.g. FPFH) andlearning-based features. Experiments on partially overlapping and noisy pointcloud registration show that our method outperforms the currentstate-of-the-art, while being more computationally efficient. Code is publiclyavailable at https://github.com/jiahaowork/idam.

Iterative Distance-Aware Similarity Matrix Convolution with Mutual-Supervised Point Elimination for Efficient Point Cloud Registration