Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks
This work addresses the problem of fast and accurate object detection in 3D point clouds for applications like autonomous driving, representing a novel method for a known bottleneck.
The paper tackled efficient object detection in 3D point clouds by proposing Vote3Deep, a method using sparse convolutional layers and L1 regularization, which outperformed previous state-of-the-art approaches by up to 40% on the KITTI benchmark while maintaining competitive processing speed.
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.