camera internal and external parameters

[duohaoxue]

Hello, model training must be heavily dependent on camera extrinsic and extrinsic parameters? Will the effect be much worse if I use it?

  I_inv = batch['intrinsics'].inverse()           # b n 3 3
  E_inv = batch['extrinsics'].inverse()           # b n 4 4

Thanks a lot!

[bradyz]

If you don't have these camera parameters the model's performance would indeed decrease - Table 3 shows the model's performance decreases by several points with unknown extrinsics.

If your camera pose is guaranteed fixed with respect to the ego vehicle, the network could learn these parameters though and still perform well

[duohaoxue]

Thank you very much for your answers