Feature Request: Mask Support for Depth Estimation

[junwon-seo716]

Hi @LiheYoung,

First, Thanks for providing the Depth-Anything code. It has been incredibly useful for estimating depth from monocular images. And It works well.

I am currently working with images captured from cameras with a wide field of view (FOV). During the calibration process to correct for distortion, certain parts of the image become black (virtual pixels). These regions do not require depth estimation, as they do not contain any useful information.

I would like to inquire if there are any plans to add mask support to your code. With mask support, we can exclude these black regions from depth estimation, which would enhance the efficiency and accuracy of the depth maps generated.

Thank you for considering this feature request. I look forward to your response.

ps. Undistorted image is like this: