Nice Work！
I have a question regarding the pose of panoramic images in the dataset. Specifically, I would like to understand the following:
How is the pose of a panoramic image defined?
Does it include both the rotation matrix and the translation matrix? If so, Is the rotation matrix of the panoramic image the same as the rotation matrix of the front view in the cubemaps?
Additionally, How is the ground truth for the panoramic image pose obtained in the dataset? Is the ground truth camera pose derived from Blender camera data or from Structure-from-Motion (SfM) methods? If it is obtained from Blender camera data, could you provide more details on how the camera pose is converted or transformed from Blender’s camera data to the format used in the dataset (e.g., rotation matrix, translation vector)?
Thank you in advance for your help!
Best regards.
Nice Work！
I have a question regarding the pose of panoramic images in the dataset. Specifically, I would like to understand the following:
How is the pose of a panoramic image defined?
Does it include both the rotation matrix and the translation matrix? If so, Is the rotation matrix of the panoramic image the same as the rotation matrix of the front view in the cubemaps?
Additionally, How is the ground truth for the panoramic image pose obtained in the dataset? Is the ground truth camera pose derived from Blender camera data or from Structure-from-Motion (SfM) methods? If it is obtained from Blender camera data, could you provide more details on how the camera pose is converted or transformed from Blender’s camera data to the format used in the dataset (e.g., rotation matrix, translation vector)?
Thank you in advance for your help!
Best regards.