Home // AISyS 2024, The First International Conference on AI-based Systems and Services // View article

Monocular Depth Estimation Pre-training for Imitation-based Autonomous Driving

Authors:
Shubham Juneja
Virginijus Marcinkevičius
Povilas Daniušis

Keywords: imitation learning; autonomous driving; monocular depth estimation; pre-training

Abstract:
Artificial intelligence based systems have taken industries and research by storm, one of such systems are employed in autonomous driving. Recent empirical findings in imitation learning for autonomous driving indicate that pre-training on various tasks can enhance the effectiveness of the learner method (e.g., neural network). We propose pre-training neural networks over the task of monocular depth estimation could be beneficial in terms of estimating another modality and extending the scene understanding capabilities of the learner method. We also outline a plan for further investigation of this approach, aiming to integrate new experimental results with existing findings in this line of research, i.e., pre-training for autonomous driving.

Pages: 35 to 37

Copyright: Copyright (c) IARIA, 2024

Publication date: September 29, 2024

Published in: conference

ISBN: 978-1-68558-192-3

Location: Venice, Italy

Dates: from September 29, 2024 to October 3, 2024