Submitted by Francis Engelmann.

Submission data

Full nameDilated Point Convolutions
DescriptionIn this work, we propose Dilated Point Convolutions (DPC). In a thorough ablation study, we show that the receptive field size is directly related to the performance of 3D point cloud processing tasks, including semantic segmentation and object classification. Point convolutions are widely used to efficiently process 3D data representations such as point clouds or graphs. However, we observe that the receptive field size of recent point convolutional networks is inherently limited. Our dilated point convolutions alleviate this issue, they significantly increase the receptive field size of point convolutions. Importantly, our dilation mechanism can easily be integrated into most existing point convolutional networks. To evaluate the resulting network architectures, we visualize the receptive field and report competitive scores on popular point cloud benchmarks.
Publication titleDilated Point Convolutions: On the Receptive Field Size of Point Convolutions on 3D Point Clouds
Publication authorsFrancis Engelmann, Theodora Kontogianni, Bastian Leibe
Publication venueICRA 2020
Publication URLhttps://francisengelmann.github.io/DPC/
Input Data TypesUses Color        Uses 3D
Programming language(s)Python and TensorFlow
HardwareGTX 1080 Ti
Websitehttps://francisengelmann.github.io/DPC/
Submission creation date8 Nov, 2018
Last edited5 Apr, 2020

3D semantic label results

Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
0.5920.7200.7000.6020.4800.7620.3800.7130.5850.4370.9400.3690.2880.4340.5090.5900.6390.5670.7720.7550.592