A Novel Neural Network Training Method for Autonomous Driving Using Semi-Pseudo-Labels and 3D Data Augmentations

TL;DR

This paper introduces a new neural network training approach for autonomous driving that combines semi-pseudo-labeling with innovative 3D data augmentations to improve 3D object detection capabilities.

Contribution

It proposes a novel training method integrating semi-pseudo-labels and 3D augmentations, enhancing detection range and accuracy in autonomous driving neural networks.

Findings

Significantly increased detection range compared to training data distribution.

Effective use of semi-pseudo-labeling for 3D object detection.

Enhanced robustness with novel 3D data augmentations.

Abstract

Training neural networks to perform 3D object detection for autonomous driving requires a large amount of diverse annotated data. However, obtaining training data with sufficient quality and quantity is expensive and sometimes impossible due to human and sensor constraints. Therefore, a novel solution is needed for extending current training methods to overcome this limitation and enable accurate 3D object detection. Our solution for the above-mentioned problem combines semi-pseudo-labeling and novel 3D augmentations. For demonstrating the applicability of the proposed method, we have designed a convolutional neural network for 3D object detection which can significantly increase the detection range in comparison with the training data distribution.