CLiNet: Joint Detection of Road Network Centerlines in 2D and 3D

TL;DR

This paper presents CLiNet, a novel method for joint detection of road network centerlines in 2D and 3D from images, utilizing a large urban driving dataset for training and evaluation.

Contribution

Introduces a new approach for joint 2D and 3D centerline detection using feature extraction tailored for urban driving tasks, with a large annotated dataset.

Findings

Achieved an F1 score of 0.684 in centerline detection.

Demonstrated effective dynamic scene modeling in urban scenarios.

Provided publicly available code and annotations.

Abstract

This work introduces a new approach for joint detection of centerlines based on image data by localizing the features jointly in 2D and 3D. In contrast to existing work that focuses on detection of visual cues, we explore feature extraction methods that are directly amenable to the urban driving task. To develop and evaluate our approach, a large urban driving dataset dubbed AV Breadcrumbs is automatically labeled by leveraging vector map representations and projective geometry to annotate over 900,000 images. Our results demonstrate potential for dynamic scene modeling across various urban driving scenarios. Our model achieves an F1 score of 0.684 and an average normalized depth error of 2.083. The code and data annotations are publicly available.