End-to-end Learning Improves Static Object Geo-localization in Monocular Video

TL;DR

This paper introduces an end-to-end learning system that jointly optimizes multiple components to improve static object geo-localization, specifically traffic lights, from monocular video in autonomous driving scenarios.

Contribution

It presents a novel joint-training approach for object pose estimation, association, and tracking, significantly enhancing localization accuracy over existing methods.

Findings

Significant performance improvements over contemporary methods.

Joint training further enhances overall system accuracy.

Effective localization of static objects like traffic lights in monocular video.

Abstract

Accurately estimating the position of static objects, such as traffic lights, from the moving camera of a self-driving car is a challenging problem. In this work, we present a system that improves the localization of static objects by jointly-optimizing the components of the system via learning. Our system is comprised of networks that perform: 1) 5DoF object pose estimation from a single image, 2) association of objects between pairs of frames, and 3) multi-object tracking to produce the final geo-localization of the static objects within the scene. We evaluate our approach using a publicly-available data set, focusing on traffic lights due to data availability. For each component, we compare against contemporary alternatives and show significantly-improved performance. We also show that the end-to-end system performance is further improved via joint-training of the constituent models.