ViewBoxTask

view.box — six orthographic faces of one region, as one image (opal-vision §OVD-6).

In the namespace opal::vision. Declared in src/libs/libopal/opal/vision/ViewBoxTask.hpp.

Detailed description

Every face is orthographic and every face shares one scale. That is what makes the six comparable: a feature 10 pixels wide on one face is the same size on the next. A perspective box would reintroduce the ambiguity the capture exists to remove.

Six faces arrive as ONE image rather than six. The cost is identical, because cost follows total pixel area — so the atlas is chosen for what it adds: the model sees six relationships at once, and each face carries a printed label, so "the spokes are wrong in -Y" means something exact.