To complete or to estimate, that is the question: A Multi-Task Depth Completion and Monocular Depth Estimation

Atapour-Abarghouei, Amir; Breckon, Toby P.

Abstract

Robust three-dimensional scene understanding is now an ever-growing area of research highly relevant in many real-world applications such as autonomous driving and robotic navigation. In this paper, we propose a multi-task learning-based model capable of performing two tasks:- sparse depth completion (i.e. generating complete dense scene depth given a sparse depth image as the input) and monocular depth estimation (i.e. predicting scene depth from a single RGB image) via two sub-networks jointly trained end to end using data randomly sampled from a publicly available corpus of synthetic and real-world images. The first sub-network generates a sparse depth image by learning lower level features from the scene and the second predicts a full dense depth image of the entire scene, leading to a better geometric and contextual understanding of the scene and, as a result, superior performance of the approach. The entire model can be used to infer complete scene depth from a single RGB image or the second network can be used alone to perform depth completion given a sparse depth input. Using adversarial training, a robust objective function, a deep architecture relying on skip connections and a blend of synthetic and real-world training data, our approach is capable of producing superior high quality scene depth. Extensive experimental evaluation demonstrates the efficacy of our approach compared to contemporary state-of-the-art techniques across both problem domains.

Citation

Atapour-Abarghouei, A., & Breckon, T. P. (2019). To complete or to estimate, that is the question: A Multi-Task Depth Completion and Monocular Depth Estimation. In Proceedings of 2019 International Conference on 3D Vision (3DV) (183-193). https://doi.org/10.1109/3dv.2019.00029

• Conference Name: International Conference on 3D Vision
• Conference Location: Quebec
• Start Date: Sep 16, 2019
• End Date: Sep 19, 2019
• Acceptance Date: Jul 30, 2019
• Publication Date: Sep 1, 2019
• Deposit Date: Aug 14, 2019
• Publicly Available Date: Nov 12, 2019
• Pages: 183-193
• Series ISSN: 2475-7888
• Book Title: Proceedings of 2019 International Conference on 3D Vision (3DV)
• DOI: https://doi.org/10.1109/3dv.2019.00029