MLCompletion: Point Cloud Completion via Multi-View Feature Augmentation and Progressive Refinement

doi:10.1109/jsen.2026.3722855

MLCompletion: Point Cloud Completion via Multi-View Feature Augmentation and Progressive Refinement

2026-08-18

0

PRE

AI

DOI：10.1109/jsen.2026.3722855

Original

Original request for help

Save

Abstract

En 中文

Point cloud completion is a fundamental task in 3D sensing and vision, yet it remains challenging when partial observations suffer from severe structural missing regions and ambiguous local geometry. To address these issues, we propose MLCompletion, a point-cloud-only completion framework that incorporates self-generated multi-view geometric augmentation and progressive shape refinement without relying on external images. Specifically, the proposed Geometry-aware Fusion Module (GFM) projects the partial point cloud into three orthogonal depth maps and fuses the extracted depth features with 3D point features, thereby enhancing global shape perception and missing-region representation. Based on this representation, the Shape Completion Module (SCP) progressively refines the coarse prediction in a coarse-to-fine manner. Within SCP, the Dynamic Feature Fusion Gate (DFFG) adaptively balances local geometric details and global structural context to improve reconstruction fidelity. Experiments on ShapeNet-55/34, PCN, and KITTI demonstrate that MLCompletion achieves competitive completion performance and consistently improves Chamfer Distance over strong baselines. Additional ablation studies, efficiency analysis, DFFG interpretability results, and extreme sparsity tests further validate the effectiveness of the proposed design while revealing its limitations under highly challenging incomplete inputs.

Keywords:

MLCompletion

Point Cloud Completion

Multi-View Augmentation

Attention Mechanism

Coarse-to-fine

MLCompletion: Point Cloud Completion via Multi-View Feature Augmentation and Progressive Refinement

Abstract

Journal

Organization

Cited Papers