eduzhai > Applied Sciences > Engineering >

GRNet Gridding Residual Network for Dense Point Cloud Completion

  • Save

... pages left unread,continue reading

Document pages: 25 pages

Abstract: Estimating the complete 3D point cloud from an incomplete one is a keyproblem in many vision and robotics applications. Mainstream methods (e.g., PCNand TopNet) use Multi-layer Perceptrons (MLPs) to directly process pointclouds, which may cause the loss of details because the structural and contextof point clouds are not fully considered. To solve this problem, we introduce3D grids as intermediate representations to regularize unordered point clouds.We therefore propose a novel Gridding Residual Network (GRNet) for point cloudcompletion. In particular, we devise two novel differentiable layers, namedGridding and Gridding Reverse, to convert between point clouds and 3D gridswithout losing structural information. We also present the differentiable CubicFeature Sampling layer to extract features of neighboring points, whichpreserves context information. In addition, we design a new loss function,namely Gridding Loss, to calculate the L1 distance between the 3D grids of thepredicted and ground truth point clouds, which is helpful to recover details.Experimental results indicate that the proposed GRNet performs favorablyagainst state-of-the-art methods on the ShapeNet, Completion3D, and KITTIbenchmarks.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×