Closed gkiavash closed 1 year ago
Multi-view refinement
3) Method Our method is able to find better matching points and therefore more accurate poses and depth maps, especially for textureless and occluded areas. At the same time, it follows the wisdom of classic methods to avoid ill-posed problems.
4) Experiments
Solve SfM problem via feature-metric bundle adjustment (BA), which explicitly enforces multi-view geometry constraints in the form of feature-metric error
Direct alignment:
Learned representation:
Metrics: Known Ground Truth:
Our approach has two main modules:
3.1. Feature and Patch Extraction
3.2. Update Operator
* : additional resourses