search

JOP-Lee / READ

AAAI2023，implementation of "READ: Large-Scale Neural Scene Rendering for Autonomous Driving", the experimental results are significantly better than Nerf-based methods
https://github.com/JOP-Lee/READ-Large-Scale-Neural-Scene-Rendering-for-Autonomous-Driving
GNU General Public License v2.0
447 stars 55 forks source link

There seems to be an error in the code #28

Open Dongber opened 1 year ago

Dongber commented 1 year ago

In line 460 of READ.gl.utils, it should be 'model['normals'] = np.zeros((n_pts, 3), dtype=np.float32)'. Is this an error?

JOP-Lee commented 1 year ago

Thanks for pointing out the error, this part of the code is using NPBG. In our work, point clouds have normals, so it doesn't matter

booker-max commented 1 year ago

Thanks for pointing out the error, this part of the code is using NPBG. In our work, point clouds have normals, so it doesn't matter

  1. When I run the pointcloud.ply file you provided, there is no normals is this file, I think you better confirm
  2. In the utils.py, you write " if model['normals'] is None: print(f'no normals in {model_path}') model['rgb'] = np.zeros((n_pts, 3), dtype=np.float32)" I think this is wrong, it should be model["normals"] = np.zeros((n_pts, 3), dtype=np.float32)", because we have already defined model["rgb"] before.