Hello, thank you for your work.
We hope to adapt SSCBench to fisheye camera for industrial work, and therefore need to obtain panoramic point clouds with semantic annotation. However, we discovered that SSCBench-KITTI-360 provides only forward-facing point clouds. Currently, the way we get labeled point clouds is by projecting point clouds onto the 2D semantic labels provided by KITTI-360. However, this approach is neither accurate nor allows for a panoramic point cloud.
Could you please share your way to get dense semantic label for panoramic point clouds?