TORNADO-Net: mulTiview tOtal vaRiatioN semAntic segmentation with Diamond inceptiOn module
Authors:
- 30 May 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 9543-9549
Abstract
Semantic segmentation of point clouds is a key component of scene understanding for robotics and autonomous driving. In this paper, we introduce TORNADO-Net - a neural network for 3D LiDAR point cloud semantic segmentation. We incorporate a multi-view (bird-eye and range) projection feature extraction with an encoder-decoder ResNet architecture with a novel diamond context block. Current projection-based methods do not take into account that neighboring points usually belong to the same class. To better utilize this local neighbourhood information and reduce noisy predictions, we introduce a combination of Total Variation, Lovász-Softmax, and Weighted Cross-Entropy losses. We also take advantage of the fact that the LiDAR data encompasses 360 ◦ field of view and use circular padding. We demonstrate state-of-the-art results on the SemanticKITTI dataset and also provide thorough quantitative evaluations and ablation results.Keywords
All Related Versions
This publication has 24 references indexed in Scilit:
- Large-Scale Point Cloud Semantic Segmentation with Superpoint Graphs2018
- SPLATNet: Sparse Lattice Networks for Point Cloud Processing2018
- Tangent Convolutions for Dense Prediction in 3D2018
- SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud2018
- Focal Loss for Dense Object Detection2017
- Multi-Task Learning and Weighted Cross-Entropy for DNN-Based Keyword Spotting2016
- Deep Residual Learning for Image Recognition2016
- Understanding deep image representations by inverting them2015
- Adaptive total variation denoising based on difference curvatureImage and Vision Computing, 2010
- Nonlinear total variation based noise removal algorithmsPhysica D: Nonlinear Phenomena, 1992