search

tsinghua-rll / VoxelNet-tensorflow

A 3D object detection system for autonomous driving.
MIT License
453 stars 123 forks source link

RPN last layer #10

Closed ZiningWang closed 6 years ago

ZiningWang commented 6 years ago

I notice you use the relu layer for the last layer of RPN which predicts class score and regression. However, as far as I know, the last layer of RPN should not use Relu. For example, if you want to shrink your anchor, you must have some negative prediction in regression.

The paper says 'BN and relu for all conv' but I doubt it is not applied at least for the last map.

jeasinema commented 6 years ago

Many thanks for this issue. Fix at jeasinema/tf_voxelnet@a7a2729.