Deep traffic light detection by overlaying synthetic context on arbitrary natural images

TL;DR

This paper introduces a synthetic data generation method using basic graphics to improve traffic light detection in autonomous driving, reducing annotation costs and addressing data imbalance, with results comparable to real data.

Contribution

It presents a novel approach to generate artificial traffic scenes on arbitrary images, enhancing training data for traffic light detection without manual annotation.

Findings

Achieved nearly 4 percentage points higher mAP than real data-based models.

Generated large synthetic datasets without annotation effort.

Improved detection of yellow traffic light states.

Abstract

Deep neural networks come as an effective solution to many problems associated with autonomous driving. By providing real image samples with traffic context to the network, the model learns to detect and classify elements of interest, such as pedestrians, traffic signs, and traffic lights. However, acquiring and annotating real data can be extremely costly in terms of time and effort. In this context, we propose a method to generate artificial traffic-related training data for deep traffic light detectors. This data is generated using basic non-realistic computer graphics to blend fake traffic scenes on top of arbitrary image backgrounds that are not related to the traffic domain. Thus, a large amount of training data can be generated without annotation efforts. Furthermore, it also tackles the intrinsic data imbalance problem in traffic light datasets, caused mainly by the low amount of samples of the yellow state. Experiments show that it is possible to achieve results comparable to those obtained with real training data from the problem domain, yielding an average mAP and an average F1-score which are each nearly 4 p.p. higher than the respective metrics obtained with a real-world reference model.