no code implementations • ICCV 2021 • Maksym Ivashechkin, Daniel Barath, Jiri Matas
Experiments on four standard datasets show that VSAC is significantly faster than all its predecessors and runs on average in 1-2 ms, on a CPU.
no code implementations • 11 Apr 2021 • Maksym Ivashechkin, Daniel Barath, Jiri Matas
We review the most recent RANSAC-like hypothesize-and-verify robust estimators.
no code implementations • CVPR 2020 • Daniel Barath, Jana Noskova, Maksym Ivashechkin, Jiri Matas
A new method for robust estimation, MAGSAC++, is proposed.
no code implementations • 5 Jun 2019 • Daniel Barath, Maksym Ivashechkin, Jiri Matas
We propose Progressive NAPSAC, P-NAPSAC in short, which merges the advantages of local and global sampling by drawing samples from gradually growing neighborhoods.
