Hey, I am trying to train HexPlane with TUM-RGBD data. In the the test results, I have weird output (the camera is go out from the scene), it looks like this: https://drive.google.com/file/d/172v2nV88qT0PHSaEhEItK5Kv30oZiYp-/view?usp=sharing. I am using droid slam's output camera trajectory. My questions is: Is the problem about predicted camera trajectory or would it be about a parameter of HexPlane ? I thought it would be about the wrong scene bbox parameters. Also the quality is pretty low, do you have any suggestion to increase it ?
Hey, I am trying to train HexPlane with TUM-RGBD data. In the the test results, I have weird output (the camera is go out from the scene), it looks like this: https://drive.google.com/file/d/172v2nV88qT0PHSaEhEItK5Kv30oZiYp-/view?usp=sharing. I am using droid slam's output camera trajectory. My questions is: Is the problem about predicted camera trajectory or would it be about a parameter of HexPlane ? I thought it would be about the wrong scene bbox parameters. Also the quality is pretty low, do you have any suggestion to increase it ?