Fix: depth unprojection

[barwojcik]

Fixed missing half-pixel shift in depth unprojection functions. The PR addresses #354.

Why was it missing?

Let's consider a 518 px x 518 px image, with height and width H, W = 518, and principal point coordinates cx, cy = W/2, H/2 = 259. In the image coordinate system (x, y with 0,0 in the top-left corner), pixel cords range from 0 to 517.
The current implementation does not perform the half-pixel shift, so after transforming to the u, v coordinate system (centered at the principal point) by subtracting cx, cy, the new pixel coordinates range from -259 to 258. That means the top-left corner has u, v = (-259, -259) and the bottom-right corner = (258, 258).
For a pinhole camera, these should be symmetrical around zero. As a result, the rays used for depth unprojection are slightly off, producing incorrect 3D points.
Performing the half-pixel shift is standard practice to correct u, v to a symmetric range from -258.5 to 258.5.

[meta-cla]

Hi @barwojcik!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at [email protected]. Thanks!

[meta-cla]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!