_version_ 1866916706428911616
author Obukhov, Anton
Poggi, Matteo
Tosi, Fabio
Arora, Ripudaman Singh
Spencer, Jaime
Russell, Chris
Hadfield, Simon
Bowden, Richard
Wang, Shuaihang
Ma, Zhenxin
Chen, Weijie
Xu, Baobei
Sun, Fengyu
Xie, Di
Zhu, Jiang
Lavreniuk, Mykola
Guan, Haining
Wu, Qun
Zeng, Yupei
Lu, Chao
Wang, Huanran
Zhou, Guangyuan
Zhang, Haotian
Wang, Jianxiong
Rao, Qiang
Wang, Chunjie
Liu, Xiao
Lou, Zhiqiang
Jiang, Hualie
Chen, Yihao
Xu, Rui
Tan, Minglang
Qin, Zihan
Mao, Yifan
Liu, Jiayang
Xu, Jialei
Yang, Yifan
Zhao, Wenbo
Jiang, Junjun
Liu, Xianming
Zhao, Mingshuai
Ming, Anlong
Chen, Wu
Xue, Feng
Yu, Mengying
Gao, Shida
Wang, Xiangfeng
Omotara, Gbenga
Farag, Ramy
Demby, Jacket
Tousi, Seyed Mohamad Ali
DeSouza, Guilherme N
Yang, Tuan-Anh
Nguyen, Minh-Quang
Tran, Thien-Phuc
Luginov, Albert
Shahzad, Muhammad
author_facet Obukhov, Anton
Poggi, Matteo
Tosi, Fabio
Arora, Ripudaman Singh
Spencer, Jaime
Russell, Chris
Hadfield, Simon
Bowden, Richard
Wang, Shuaihang
Ma, Zhenxin
Chen, Weijie
Xu, Baobei
Sun, Fengyu
Xie, Di
Zhu, Jiang
Lavreniuk, Mykola
Guan, Haining
Wu, Qun
Zeng, Yupei
Lu, Chao
Wang, Huanran
Zhou, Guangyuan
Zhang, Haotian
Wang, Jianxiong
Rao, Qiang
Wang, Chunjie
Liu, Xiao
Lou, Zhiqiang
Jiang, Hualie
Chen, Yihao
Xu, Rui
Tan, Minglang
Qin, Zihan
Mao, Yifan
Liu, Jiayang
Xu, Jialei
Yang, Yifan
Zhao, Wenbo
Jiang, Junjun
Liu, Xianming
Zhao, Mingshuai
Ming, Anlong
Chen, Wu
Xue, Feng
Yu, Mengying
Gao, Shida
Wang, Xiangfeng
Omotara, Gbenga
Farag, Ramy
Demby, Jacket
Tousi, Seyed Mohamad Ali
DeSouza, Guilherme N
Yang, Tuan-Anh
Nguyen, Minh-Quang
Tran, Thien-Phuc
Luginov, Albert
Shahzad, Muhammad
contents This paper presents the results of the fourth edition of the Monocular Depth Estimation Challenge (MDEC), which focuses on zero-shot generalization to the SYNS-Patches benchmark, a dataset featuring challenging environments in both natural and indoor settings. In this edition, we revised the evaluation protocol to use least-squares alignment with two degrees of freedom to support disparity and affine-invariant predictions. We also revised the baselines and included popular off-the-shelf methods: Depth Anything v2 and Marigold. The challenge received a total of 24 submissions that outperformed the baselines on the test set; 10 of these included a report describing their approach, with most leading methods relying on affine-invariant predictions. The challenge winners improved the 3D F-Score over the previous edition's best result, raising it from 22.58% to 23.05%.
format Preprint
id arxiv_https___arxiv_org_abs_2504_17787
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle The Fourth Monocular Depth Estimation Challenge
Obukhov, Anton
Poggi, Matteo
Tosi, Fabio
Arora, Ripudaman Singh
Spencer, Jaime
Russell, Chris
Hadfield, Simon
Bowden, Richard
Wang, Shuaihang
Ma, Zhenxin
Chen, Weijie
Xu, Baobei
Sun, Fengyu
Xie, Di
Zhu, Jiang
Lavreniuk, Mykola
Guan, Haining
Wu, Qun
Zeng, Yupei
Lu, Chao
Wang, Huanran
Zhou, Guangyuan
Zhang, Haotian
Wang, Jianxiong
Rao, Qiang
Wang, Chunjie
Liu, Xiao
Lou, Zhiqiang
Jiang, Hualie
Chen, Yihao
Xu, Rui
Tan, Minglang
Qin, Zihan
Mao, Yifan
Liu, Jiayang
Xu, Jialei
Yang, Yifan
Zhao, Wenbo
Jiang, Junjun
Liu, Xianming
Zhao, Mingshuai
Ming, Anlong
Chen, Wu
Xue, Feng
Yu, Mengying
Gao, Shida
Wang, Xiangfeng
Omotara, Gbenga
Farag, Ramy
Demby, Jacket
Tousi, Seyed Mohamad Ali
DeSouza, Guilherme N
Yang, Tuan-Anh
Nguyen, Minh-Quang
Tran, Thien-Phuc
Luginov, Albert
Shahzad, Muhammad
Computer Vision and Pattern Recognition
This paper presents the results of the fourth edition of the Monocular Depth Estimation Challenge (MDEC), which focuses on zero-shot generalization to the SYNS-Patches benchmark, a dataset featuring challenging environments in both natural and indoor settings. In this edition, we revised the evaluation protocol to use least-squares alignment with two degrees of freedom to support disparity and affine-invariant predictions. We also revised the baselines and included popular off-the-shelf methods: Depth Anything v2 and Marigold. The challenge received a total of 24 submissions that outperformed the baselines on the test set; 10 of these included a report describing their approach, with most leading methods relying on affine-invariant predictions. The challenge winners improved the 3D F-Score over the previous edition's best result, raising it from 22.58% to 23.05%.
title The Fourth Monocular Depth Estimation Challenge
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.17787