This paper discusses the results of the third edition of the Monocular Depth Estimation Challenge (MDEC). The challenge focuses on zero-shot generalization to the challenging SYNS-Patches dataset, featuring complex scenes in natural and indoor settings. As with the previous edition, methods can use any form of supervision, i.e. supervised or self-supervised. The challenge received a total of 19 submissions outperforming the baseline on the test set: 10 among them submitted a report describing their approach, highlighting a diffused use of foundational models such as Depth Anything at the core of their method. The challenge winners drastically improved 3D F-Score performance, from 17.51% to 23.72%.
@article{arxiv.2404.16831,
title = {The Third Monocular Depth Estimation Challenge},
author = {Jaime Spencer and Fabio Tosi and Matteo Poggi and Ripudaman Singh Arora and Chris Russell and Simon Hadfield and Richard Bowden and GuangYuan Zhou and ZhengXin Li and Qiang Rao and YiPing Bao and Xiao Liu and Dohyeong Kim and Jinseong Kim and Myunghyun Kim and Mykola Lavreniuk and Rui Li and Qing Mao and Jiang Wu and Yu Zhu and Jinqiu Sun and Yanning Zhang and Suraj Patni and Aradhye Agarwal and Chetan Arora and Pihai Sun and Kui Jiang and Gang Wu and Jian Liu and Xianming Liu and Junjun Jiang and Xidan Zhang and Jianing Wei and Fangjun Wang and Zhiming Tan and Jiabao Wang and Albert Luginov and Muhammad Shahzad and Seyed Hosseini and Aleksander Trajcevski and James H. Elder},
journal= {arXiv preprint arXiv:2404.16831},
year = {2024}
}