Lidar-Camera Co-Training for Semi-Supervised Road Detection

TL;DR

This paper demonstrates that co-training, a semi-supervised learning approach, significantly improves road detection accuracy for lidar and camera systems by effectively leveraging unlabeled data, reducing the need for extensive manual labeling.

Contribution

It introduces a co-training method for semi-supervised road detection that enhances performance with minimal labeled data, validated on the KITTI benchmark.

Findings

F1-score improvements of up to 8.14 percentage points

Effective use of unlabeled data reduces labeling effort

High performance achieved with only 36 labeled examples

Abstract

Recent advances in the field of machine learning and computer vision have enabled the development of fast and accurate road detectors. Commonly such systems are trained within a supervised learning paradigm where both an input sensor's data and the corresponding ground truth label must be provided. The task of generating labels is commonly carried out by human annotators and it is notoriously time consuming and expensive. In this work, it is shown that a semi-supervised approach known as co-training can provide significant F1-score average improvements compared to supervised learning. In co-training, two classifiers acting on different views of the data cooperatively improve each other's performance by leveraging unlabeled examples. Depending on the amount of labeled data used, the improvements ranged from 1.12 to 6.10 percentage points for a camera-based road detector and from 1.04 to 8.14 percentage points for a lidar-based road detector. Lastly, the co-training algorithm is validated on the KITTI road benchmark, achieving high performance using only 36 labeled training examples together with several thousands unlabeled ones.