do you save keypoints through frames ?

[an99990]

In your paper, it is mentioned that you do tracking, however I am unable to find where in the source code you keep track of the features detected.
Is it possible to refer me to where the tracking is happening ? how do you keep the features over time ?

Thank you

[ymd-stella]

In your paper

There is a misunderstanding. The paper is a copyrighted work of AIST. I am not a member of AIST, hence the name change to OpenVSLAM. Hence the name change from OpenVSLAM to stella_vslam, see #267.

it is mentioned that you do tracking, however I am unable to find where in the source code you keep track of the features detected.
Is it possible to refer me to where the tracking is happening ? how do you keep the features over time ?

Feature points are not directly tracked. Landmarks are generated from the feature points and the camera pose and landmarks are tracked. See

stella_vslam/src/stella_vslam/tracking_module.cc

Line 200 in e6d8fa5

bool tracking_module::track(bool relocalization_is_needed) {