no code implementations • 16 Jan 2019 • Shridhar Ravikumar
We quantitatively evaluate our results comparing it with an approach using a monocular input without our spatial constraints and show that our results are closer to the ground-truth geometry.
