Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zeng, Yadan, Zhou, Jiadong, Li, Xiaohan, Chen, I-Ming
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2509.13692
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908543731367936
author Zeng, Yadan
Zhou, Jiadong
Li, Xiaohan
Chen, I-Ming
author_facet Zeng, Yadan
Zhou, Jiadong
Li, Xiaohan
Chen, I-Ming
contents Point cloud completion is essential for robotic perception, object reconstruction and supporting downstream tasks like grasp planning, obstacle avoidance, and manipulation. However, incomplete geometry caused by self-occlusion and sensor limitations can significantly degrade downstream reasoning and interaction. To address these challenges, we propose HGACNet, a novel framework that reconstructs complete point clouds of individual objects by hierarchically encoding 3D geometric features and fusing them with image-guided priors from a single-view RGB image. At the core of our approach, the Hierarchical Graph Attention (HGA) encoder adaptively selects critical local points through graph attention-based downsampling and progressively refines hierarchical geometric features to better capture structural continuity and spatial relationships. To strengthen cross-modal interaction, we further design a Multi-Scale Cross-Modal Fusion (MSCF) module that performs attention-based feature alignment between hierarchical geometric features and structured visual representations, enabling fine-grained semantic guidance for completion. In addition, we proposed the contrastive loss (C-Loss) to explicitly align the feature distributions across modalities, improving completion fidelity under modality discrepancy. Finally, extensive experiments conducted on both the ShapeNet-ViPC benchmark and the YCB-Complete dataset confirm the effectiveness of HGACNet, demonstrating state-of-the-art performance as well as strong applicability in real-world robotic manipulation tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2509_13692
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle HGACNet: Hierarchical Graph Attention Network for Cross-Modal Point Cloud Completion
Zeng, Yadan
Zhou, Jiadong
Li, Xiaohan
Chen, I-Ming
Robotics
Point cloud completion is essential for robotic perception, object reconstruction and supporting downstream tasks like grasp planning, obstacle avoidance, and manipulation. However, incomplete geometry caused by self-occlusion and sensor limitations can significantly degrade downstream reasoning and interaction. To address these challenges, we propose HGACNet, a novel framework that reconstructs complete point clouds of individual objects by hierarchically encoding 3D geometric features and fusing them with image-guided priors from a single-view RGB image. At the core of our approach, the Hierarchical Graph Attention (HGA) encoder adaptively selects critical local points through graph attention-based downsampling and progressively refines hierarchical geometric features to better capture structural continuity and spatial relationships. To strengthen cross-modal interaction, we further design a Multi-Scale Cross-Modal Fusion (MSCF) module that performs attention-based feature alignment between hierarchical geometric features and structured visual representations, enabling fine-grained semantic guidance for completion. In addition, we proposed the contrastive loss (C-Loss) to explicitly align the feature distributions across modalities, improving completion fidelity under modality discrepancy. Finally, extensive experiments conducted on both the ShapeNet-ViPC benchmark and the YCB-Complete dataset confirm the effectiveness of HGACNet, demonstrating state-of-the-art performance as well as strong applicability in real-world robotic manipulation tasks.
title HGACNet: Hierarchical Graph Attention Network for Cross-Modal Point Cloud Completion
topic Robotics
url https://arxiv.org/abs/2509.13692