Efficient Registration of Multi-View Point Sets by K-means Clustering

This paper casts multi-view registration as a clustering task that can be solved using an approach proposed based on K-means clustering. For clustering, all centroids are uniformly sampled from the initially aligned point sets involved in multi-view registration, to render clustering efficient and effective. Then, each point is assigned to a single cluster, and each cluster centroid is updated accordingly. Subsequently, the shape constituted by all cluster centroids is used to estimate the rigid transformation for each point set sequentially. For accuracy and stability, clustering and transformation estimation are alternately and iteratively applied to all point sets. We tested our proposed approach on several benchmark datasets and compared it with state-of-the-art approaches. Experimental results validate its efficiency and robustness for the registration of multi-view point sets.