Free-Space Detection with Self-Supervised and Online Trained Fully Convolutional Networks

TL;DR

This paper presents a self-supervised, online-trained Fully Convolutional Network for free-space detection in autonomous driving, reducing reliance on manual labels and improving adaptability and performance in diverse traffic scenes.

Contribution

It introduces a self-supervised training method using stereo disparity for free-space detection, enabling online training and adaptation of FCNs in real-time driving scenarios.

Findings

Online training improves performance by 5% over offline training.

Self-supervised approach achieves comparable results to manual annotation.

Validated on public and new challenging datasets.

Abstract

Recently, vision-based Advanced Driver Assist Systems have gained broad interest. In this work, we investigate free-space detection, for which we propose to employ a Fully Convolutional Network (FCN). We show that this FCN can be trained in a self-supervised manner and achieve similar results compared to training on manually annotated data, thereby reducing the need for large manually annotated training sets. To this end, our self-supervised training relies on a stereo-vision disparity system, to automatically generate (weak) training labels for the color-based FCN. Additionally, our self-supervised training facilitates online training of the FCN instead of offline. Consequently, given that the applied FCN is relatively small, the free-space analysis becomes highly adaptive to any traffic scene that the vehicle encounters. We have validated our algorithm using publicly available data and on a new challenging benchmark dataset that is released with this paper. Experiments show that the online training boosts performance with 5% when compared to offline training, both for Fmax and AP.