We propose LU-Net (for LiDAR U-Net), for the semantic segmentation of a 3D LiDAR point cloud. Instead of applying some global 3D segmentation method such as Point-Net, we propose an end-to-end architecture for LiDAR point cloud semantic segmentation that efficiently solves the problem as an image processing problem. First, a high-level 3D feature extraction module is used to compute 3D local features for each point given its neighbors. Then, these features are projected into a 2D multichannel range-image by considering the topology of the sensor. This range-image later serves as the input to a U-Net segmentation network, which is a simple architecture yet enough for our purpose. In this way, we can exploit both the 3D nature of the data and the specificity of the LiDAR sensor. This approach efficiently bridges between 3D point cloud processing and image processing as it outperforms the state-of-the-art by a large margin on the KITTI dataset, as our experiments show. Moreover, this approach operates at 24fps on a single GPU. This is above the acquisition rate of common LiDAR sensors which makes it suitable for real-time applications.
|Publication status||Published - 2019|
|Event||IEEE International Conference on Computer Vision: ICCV 2019 - Seoul, Seoul, Korea, Republic of|
Duration: 27 Oct 2019 → 2 Nov 2019
|Conference||IEEE International Conference on Computer Vision|
|Abbreviated title||ICCV 2019|
|Country/Territory||Korea, Republic of|
|Period||27/10/19 → 2/11/19|