Description: “We’re all stories, in the end. Just make it a good one, eh?” — The Doctor, Season 5, Episode 13
Configurations of my hexo blog, now rendering with pandoc and automatic building/publishing with Travis-CI.
Camera pose estimation is one of the most widely used low-level computer vision research, fundamentally supports SLAM, SfM, AR, VR and our ACR (Active Camera Relocalization).
For one camera, we can express its pose by its location and rotation w.r.t. to one global coordinate, and for relative pose between two images, we can treat one as the global and assign \(I=[eye(3);0]\) to it. But this kind of intuitive description/assignment is not clear enough at all, especially when you try to code it (believe me -.-).