CenFormer: Transformer-based Network from Centroid Generation for Point Cloud Completion

Point clouds captured from 3D scanners are often sparse and incomplete due to occlusions, limited viewpoints, and sensor constraints. These limitations hinder applications in robotics, autonomous navigation, and augmented reality. Hence, point cloud completion is crucial for generating reliable 3D object representations. Existing methods often struggle to capture structural patterns effectively, which leads to low-quality reconstructions. To address these challenges, we propose Centroid Transformer (CenFormer), a novel transformer-based network for point cloud completion. CenFormer introduces two distinct types of centroids, namely Preserved and Dispersed, to facilitate fine-grained reconstruction. The proposed design includes three innovations: i) a Centroid Generation Block to aggregate features for preserved centroids, ii) a Centroid Dispersion Block to predict offsets for dispersed centroids, and iii) a Fine-grained Point Generation Block to refine local patterns around centroids. These components jointly enable the network to effectively capture local structural details and strategically target missing regions for fine-grained 3D shape reconstruction. Experiments on various benchmark datasets demonstrate that CenFormer significantly outperforms state-of-the-art methods in both visualization results and quantitative metrics.