SEGO was a system developed in 2018 with my collaborators Victor Lempitsky and Yinqiang Zheng. We studied a problem of robust pose alignment between two stereo cameras based on line and point features. We arrived at a new algorithm which uses minimal possible number of observations to solve the problem. The method can significantly increase robustess of the well-known ORB-SLAM2 system. Please see the project page here.