Erick Mendieta

Results 7 comments of Erick Mendieta

Thank you for your reply. In this case, every position is a translation of the VLN dataset which corresponds to the position of the cameras in the environment. **However, the...

Unfortunately, I have compared both locations, and we have the same In all cases. The PR removed 3220 locations with this problem from the dataset. (I checked 3-5 and there...

I see that makes sense. Then I understand the problem now. I will try to code around it. Thank you for the explanation.

@mathfac Thank you for your answer. I want to ask you about the calculation of the z value translated from matterport to habitat, Did you took the information on the...

tl:dr, in order to translate matterport points to habitat in addition to the code mention by @erikwijmans you need to subtract the camera z value with respect to the agent...

@mathfac yes I follow your advice to use the snap point. So I did the following: 1. First I tried to calculate the geodesic distance, between start_point and goal. 2....

@erikwijmans Thank you again for your comments really valuable indeed!. And yes, you are right the Matterport3D camera settings are really inconsistent. However, the actions MOVE_FORWARD, TURN_LEFT, TURN_RIGHT..., automatically corrects...