Xueqian Li

Hi, we have preliminarily tried using positional encoding, which did not help in our case. My understanding is that the scene flow representation for autonomous driving scenes might not necessarily...

Hi @1404561326521, you could find a function [get_datasets](https://github.com/Lilac-Lee/PointNetLK_Revisited/blob/main/train.py#L154) in the train.py file that loads training data from your dataset folder. In the data_utils.py file, you could find a class named...

Hey, thanks for reaching out. These are answers to your questions. 1. Yes, as we also claimed in the paper, we used 10 iterations during training on the ModelNet dataset....

Hi, thanks for your comment. I have updated the code.

Hi @StrivedTye, you could consider the ReLU-MLP as ReLU(Ax+b) (see eq. 8). Therefore, you are computing the partial derivative of the network of the input. If your input has size...

Hey @praj441, you can try: 1. finding the overlapped area between the point clouds, we only need to compute the feature for overlapped areas; 2. after finding the overlapping area,...

Hi @Jarrome, zero_mean is required by our method. The voxelization part is the same analog as the synthetic part. voxel_mean is required when computing global Jacobian. Our method solves both...

Hi @rajat-talak, thanks for reaching out. What is the number of points you used for the ModelNet dataset? Also, when computing Jacobian, we only used 100 points, could you double...

Hi, I found the previous discussion about a similar issue here [https://github.com/Lilac-Lee/PointNetLK_Revisited/issues/10](https://github.com/Lilac-Lee/PointNetLK_Revisited/issues/10) is interesting, and might be helpful. Could you take a look? Cheers.

Hi @waveleaf27, this might happen when the grid cell size is relatively large, making the prediction prone to similar motions nearby. Please refer to [MBNSF](https://github.com/kavisha725/MBNSF) for the implementation of NSFP++...