I have a problem with the kind A * x = y, where I have A and y, and I want to find x. The vector x has dimension 3, and I can get up to 110 measurements to build A andy. Currently, I am solving this using RANSAC available on the scikit-learn library. However, I would like to do the same with MAGSAC, which is state-of-the-art for such tasks. I tried to implement such an estimator by myself using the 2D Line example, but I couldn't find a solution.
Is it possible to solve such a problem using the public MAGSAC code? If so, can you explain where/how to modify the files?
Hi @danini and @ducha-aiki ,
I have a problem with the kind
A * x = y
, where I have A and y, and I want to find x. The vector x has dimension 3, and I can get up to 110 measurements to buildA
andy.
Currently, I am solving this using RANSAC available on the scikit-learn library. However, I would like to do the same with MAGSAC, which is state-of-the-art for such tasks. I tried to implement such an estimator by myself using the 2D Line example, but I couldn't find a solution.Is it possible to solve such a problem using the public MAGSAC code? If so, can you explain where/how to modify the files?
Thank you, Paulo Araujo