Скачать или смотреть 3rd Monocular Depth Estimation Challenge CVPR 2024

The 3rd Monocular Depth Estimation Challenge was organised as part of a workshop at CVPR2024. The objective was to evaluate monocular depth estimation on the challenging SYNS-Patches dataset. This dataset contains complex natural, urban and indoor scenes. The first edition of the challenge was focused on evaluating only self-supervised approaches to MDE. This edition expands the focus to any existing monocular approach, including supervised, weakly/self-supervised, multi-task. The only requirement is that the method uses only a single image when making the final depth prediction. Once the challenge is complete, participants must provide details about their submission (including type of supervision) such that they can be ranked fairly.