When you calculate the error of the estimated depth map and the accuracy of <x mm, why divide by depth_interval?

[zhuixunforever]

In table 2 in your paper, you provide validation accuracy against ground truth depth maps with different models.

and I notice you divide depth_interval when calculating l1loss and less_one_accuracy.

I'm confused, can you give me some advice, because I don't think we need to divide by depth_interval here.