Convolutional Bayesian Kernel Inference for 3D Semantic Mapping

TL;DR

This paper introduces ConvBKI, a convolutional Bayesian inference layer that enhances real-time 3D semantic mapping by combining efficiency and reliability, demonstrated on the KITTI dataset.

Contribution

The paper presents a novel ConvBKI layer that performs explicit Bayesian inference within convolutional layers for improved 3D semantic mapping.

Findings

Improved latency over state-of-the-art methods

Maintained comparable semantic inference accuracy

Effective integration of semantic predictions into global maps

Abstract

Robotic perception is currently at a cross-roads between modern methods, which operate in an efficient latent space, and classical methods, which are mathematically founded and provide interpretable, trustworthy results. In this paper, we introduce a Convolutional Bayesian Kernel Inference (ConvBKI) layer which learns to perform explicit Bayesian inference within a depthwise separable convolution layer to maximize efficency while maintaining reliability simultaneously. We apply our layer to the task of real-time 3D semantic mapping, where we learn semantic-geometric probability distributions for LiDAR sensor information and incorporate semantic predictions into a global map. We evaluate our network against state-of-the-art semantic mapping algorithms on the KITTI data set, demonstrating improved latency with comparable semantic label inference results.