Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Consensus Pruning for Translation Averaging
Translation averaging is the problem of estimating the global 3D translation of a set of cameras given their orientations and a set of relative translations estimated from image pairs. Translation averaging is a cornerstone in global Structure-from-Motion algorithms. This project is about designing a deep network that iteratively removes incorrect relative translations to estimate the global ones.