Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks

TL;DR

Vote3Deep introduces a fast, sparse convolutional neural network method for 3D object detection in point clouds, leveraging a novel voting scheme and L1 regularization to improve efficiency and accuracy.

Contribution

This work is the first to implement sparse convolutional layers with L1 regularization for large-scale 3D data processing, achieving state-of-the-art results.

Findings

Outperforms previous methods by up to 40% on KITTI benchmark.

Uses fewer layers to achieve high accuracy.

Maintains high processing speed with sparse convolutions.

Abstract

This paper proposes a computationally efficient approach to detecting objects natively in 3D point clouds using convolutional neural networks (CNNs). In particular, this is achieved by leveraging a feature-centric voting scheme to implement novel convolutional layers which explicitly exploit the sparsity encountered in the input. To this end, we examine the trade-off between accuracy and speed for different architectures and additionally propose to use an L1 penalty on the filter activations to further encourage sparsity in the intermediate representations. To the best of our knowledge, this is the first work to propose sparse convolutional layers and L1 regularisation for efficient large-scale processing of 3D data. We demonstrate the efficacy of our approach on the KITTI object detection benchmark and show that Vote3Deep models with as few as three layers outperform the previous state of the art in both laser and laser-vision based approaches by margins of up to 40% while remaining highly competitive in terms of processing time.