TY - JOUR
T1 - A Convolutional Neural Network for Point Cloud Instance Segmentation in Cluttered Scene Trained by Synthetic Data without Color
AU - Xu, Yajun
AU - Arai, Shogo
AU - Tokuda, Fuyuki
AU - Kosuge, Kazuhiro
N1 - Publisher Copyright:
© 2013 IEEE.
PY - 2020
Y1 - 2020
N2 - 3D Instance segmentation is a fundamental task in computer vision. Effective segmentation plays an important role in robotic tasks, augmented reality, autonomous driving, etc. With the ascendancy of convolutional neural networks in 2D image processing, the use of deep learning methods to segment 3D point clouds receives much attention. A great convergence of training loss often requires a large amount of human-annotated data, while making such a 3D dataset is time-consuming. This paper proposes a method for training convolutional neural networks to predict instance segmentation results using synthetic data. The proposed method is based on the SGPN framework. We replaced the original feature extractor with 'dynamic graph convolutional neural networks' that learned how to extract local geometric features and proposed a simple and effective loss function, making the network more focused on hard examples. We experimentally proved that the proposed method significantly outperforms the state-of-the-art method in both Stanford 3D Indoor Semantics Dataset and our datasets.
AB - 3D Instance segmentation is a fundamental task in computer vision. Effective segmentation plays an important role in robotic tasks, augmented reality, autonomous driving, etc. With the ascendancy of convolutional neural networks in 2D image processing, the use of deep learning methods to segment 3D point clouds receives much attention. A great convergence of training loss often requires a large amount of human-annotated data, while making such a 3D dataset is time-consuming. This paper proposes a method for training convolutional neural networks to predict instance segmentation results using synthetic data. The proposed method is based on the SGPN framework. We replaced the original feature extractor with 'dynamic graph convolutional neural networks' that learned how to extract local geometric features and proposed a simple and effective loss function, making the network more focused on hard examples. We experimentally proved that the proposed method significantly outperforms the state-of-the-art method in both Stanford 3D Indoor Semantics Dataset and our datasets.
KW - Point cloud
KW - deep learning
KW - instance segmentation
UR - http://www.scopus.com/inward/record.url?scp=85083899191&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85083899191&partnerID=8YFLogxK
U2 - 10.1109/ACCESS.2020.2978506
DO - 10.1109/ACCESS.2020.2978506
M3 - Article
AN - SCOPUS:85083899191
VL - 8
SP - 70262
EP - 70269
JO - IEEE Access
JF - IEEE Access
SN - 2169-3536
M1 - 9025047
ER -