
Kasra Khosoussi
@kasrakhosoussi
Roboticist β’ EECS @UQ_News β’ ex @CSIRO, @MITLIDS @MITAeroAstro
ID: 1144062638701404161
https://www.mit.edu/~mrrobot 27-06-2019 01:59:23
319 Tweet
1,1K Followers
1,1K Following

Iman Shames As we suspected, this approach seems to work well with robust cost functions (M-estimators) like Huber. In this example, I've added outlier edges to the Intel Research Lab dataset (long red edges). The GIF on the left shows the results over 10 (outer) iterations without

Liping Bai η½θθ BTW these slides might be useful. This is from a presentation I made with Hamid Taghirad in 2013 (some of the slides come from my undergrad thesis which was full of memes - we had to take them outπ). I took a quick look and it looks accurate enough. mit.edu/~mrrobot/assetβ¦

People should start naming large models after Von Neumann for the ironyπ βI remember my friend Johnny von Neumann used to say, with four parameters I can fit an elephant, and with five I can make him wiggle his trunk.β en.wikipedia.org/wiki/Von_Neumaβ¦ Full context (A meeting with








VSLAM-LAB: A Comprehensive Framework for Visual SLAM Methods and Datasets Alejandro FontΓ‘n, Tobias Fischer, Javier Civera, Michael Milford, FTSE π€πππ§ ππ£οΈ tl;dr: in title arxiv.org/abs/2504.04457




