1 paper · 1 filter
Gwangtak Bae, Changwoon Choi, Hyeongjun Heo +2
We present an inverse image-formation module that can enhance the robustness of existing visual SLAM pipelines for casually captured scenarios. Casual video captures often suffer f…