This repo accompanies the research paper, ARKitScenes - A Diverse Real-World Dataset for 3D Indoor Scene Understanding Using Mobile RGB-D Data and contains the data, scripts to visualize and process assets, and training code described in our paper.
I use the following computation to get the position of the 2D bounding box, but it doesn't match the object in the image.
img from lower_wide[time_id]
intrinsics = lower_wide_intrinsics[time_id]
extrinsics from lowres_wide.traj
corners from the data_prepare_offline.py
I wonder whether the intrinsic matrix divides 1920/256(7.5) as a low resolution. And do I miss something operation?
I use the following computation to get the position of the 2D bounding box, but it doesn't match the object in the image. img from lower_wide[time_id] intrinsics = lower_wide_intrinsics[time_id] extrinsics from lowres_wide.traj corners from the data_prepare_offline.py
I wonder whether the intrinsic matrix divides 1920/256(7.5) as a low resolution. And do I miss something operation?