arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2501.02464cs.CVcs.AIcs.RO
收藏

Depth Any Camera:来自任意相机的零样本度量深度估计

Depth Any Camera: Zero-Shot Metric Depth Estimation from Any Camera

Bosch Research North America(博世北美研究院);Bosch Center for Artificial Intelligence (BCAI)(博世人工智能中心);Carnegie Mellon University(卡内基梅隆大学);Simon Fraser University(西蒙菲莎大学)

Yuliang Guo, Sparsh Garg, S. Mahdi H. Miangoleh, Xinyu Huang, Liu Ren

AI总结:

本文提出DAC框架,仅用透视图像训练并以ERP统一表示任意FoV图像,通过俯仰感知转换、FoV对齐和多分辨率增强实现鱼眼与360度相机的零样本度量深度SOTA。

摘要:

尽管近期的深度基础模型展现出强大的零样本泛化能力,但在不同相机类型(尤其是鱼眼和360度相机等具有大视场(FoV)的相机)上实现准确的度量深度估计仍是一项重大挑战。本文提出 Depth Any Camera(DAC),这是一个强大的零样本度量深度估计框架,将基于透视图像训练的模型扩展为能够有效处理具有不同FoV的相机。该框架旨在确保能够利用所有现有3D数据,而不受新应用中具体相机类型的限制。值得注意的是,DAC仅在透视图像上训练,却能无缝泛化到鱼眼和360度相机,无需专门的训练数据。DAC采用等矩形投影(Equi-Rectangular Projection,ERP)作为统一图像表示,从而能够一致地处理具有不同FoV的图像。其核心组件包括:俯仰感知的图像到ERP转换,配合高效的在线增强,以从未失真输入模拟失真的ERP块;FoV对齐操作,以支持在广泛FoV范围内进行有效训练;以及多分辨率数据增强,进一步解决训练与测试之间的分辨率差异。DAC实现了最先进的零样本度量深度估计;与先前的度量深度基础模型相比,在多个鱼眼和360度数据集上将 δ₁ 准确率最高提升50%,展示出跨相机类型的稳健泛化能力。

英文摘要:

While recent depth foundation models exhibit strong zero-shot generalization, achieving accurate metric depth across diverse camera types-particularly those with large fields of view (FoV) such as fisheye and 360-degree cameras-remains a significant challenge. This paper presents Depth Any Camera (DAC), a powerful zero-shot metric depth estimation framework that extends a perspective-trained model to effectively handle cameras with varying FoVs. The framework is designed to ensure that all existing 3D data can be leveraged, regardless of the specific camera types used in new applications. Remarkably, DAC is trained exclusively on perspective images but generalizes seamlessly to fisheye and 360-degree cameras without the need for specialized training data. DAC employs Equi-Rectangular Projection (ERP) as a unified image representation, enabling consistent processing of images with diverse FoVs. Its core components include pitch-aware Image-to-ERP conversion with efficient online augmentation to simulate distorted ERP patches from undistorted inputs, FoV alignment operations to enable effective training across a wide range of FoVs, and multi-resolution data augmentation to further address resolution disparities between training and testing. DAC achieves state-of-the-art zero-shot metric depth estimation, improving $δ_1$ accuracy by up to 50% on multiple fisheye and 360-degree datasets compared to prior metric depth foundation models, demonstrating robust generalization across camera types.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑