AGNet: An Attention-Based Graph Network for Point Cloud Classification and Segmentation

Remote Sensing

卷:14

Jian Wang¹  Donglin Di²  Linhui Li³  Weipeng Jing³  Wenjun Zhang³  Guangsheng Chen³ 

[1] Aerospace Information Research Institute, CAS, Beijing 100094, China;

[2] Baidu Company Ltd., Beijing 100085, China;

[3] College of Information and Computer Engineering, Northeast Forestry University, Harbin 150040, China;

关键词: geometric features;    3D point clouds;    shape analysis;    neural network;    graph attention mechanism;   

DOI  :  10.3390/rs14041036

来源: DOAJ

【 摘 要 】

Classification and segmentation of point clouds have attracted increasing attention in recent years. On the one hand, it is difficult to extract local features with geometric information. On the other hand, how to select more important features correctly also brings challenges to the research. Therefore, the main challenge in classifying and segmenting the point clouds is how to locate the attentional region. To tackle this challenge, we propose a graph-based neural network with an attention pooling strategy (AGNet). In particular, local feature information can be extracted by constructing a topological structure. Compared to existing methods, AGNet can better extract the spatial information with different distances, and the attentional pooling strategy is capable of selecting the most important features of the topological structure. Therefore, our model can aggregate more information to better represent different point cloud features. We conducted extensive experiments on challenging benchmark datasets including ModelNet40 for object classification, as well as ShapeNet Part and S3DIS for segmentation. Both the quantitative and qualitative experiments demonstrated a consistent advantage for the tasks of point set classification and segmentation.

【 授权许可】

Unknown