For the technically inclined, a slide show, a poster (in two versions) and two scientific reports about parts of CeX3D Inverse's technology have been published:
You might also want to check out: Announcement: Demo of CeX3D Inverse and Standard ML technology at ML Workshop 2012 in Copenhagen 13th of September.
CeX3D Inverse Pipeline in Layman's Terms
CeX3D Inverse starts from input pictures of the same scene, e.g. taken with a camera:
Features between images are matched and outliers (black boxes) are removed, while putative inliers (white boxes) are kept:
Camera relationships between images are established, as witnessed by the epipolar line correspondences:
The above is only the first part of the pipeline; there is a lot more to it than this.
Only information relating to the above has been presented to the public and this is already somewhat outdated.
CeX3D Inverse's critical TODO-list before alpha release (click here)
may give a few additional hints about parts of the pipeline.